From 9a909f1cf63090ee1fb65d70668c850e6df18d0a Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Thu, 14 Dec 2023 16:45:12 +0800
Subject: [PATCH 1/8] =?UTF-8?q?perf:=20=E4=BC=98=E5=8C=96=E8=BF=94?=
 =?UTF-8?q?=E5=9B=9E=E9=94=99=E8=AF=AF=E4=BF=A1=E6=81=AF?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 src/openai.js     | 6 +++++-
 src/workers-ai.js | 6 +++++-
 2 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/src/openai.js b/src/openai.js
index 797b8ed7..9feb5740 100644
--- a/src/openai.js
+++ b/src/openai.js
@@ -120,7 +120,11 @@ export async function requestCompletionsFromOpenAI(message, history, context, on
     }
   }
   setTimeout(() => updateBotUsage(result.usage, context).catch(console.error), 0);
-  return result.choices[0].message.content;
+  try {
+    return result.choices[0].message.content;
+  } catch (e) {
+    return  result?.error?.message || JSON.stringify(result);
+  }
 }
 
 
diff --git a/src/workers-ai.js b/src/workers-ai.js
index d24607ab..304b5d95 100644
--- a/src/workers-ai.js
+++ b/src/workers-ai.js
@@ -77,7 +77,11 @@ export async function requestCompletionsFromWorkersAI(message, history, context,
     return contentFull;
   } else {
     const data = await resp.json();
-    return data.result.response;
+    try {
+      return data.result.response;
+    } catch (e) {
+      return data.errors?.[0]?.message || JSON.stringify(data)
+    }
   }
 }
 

From 6d666fd73e6dba3b603cbd2071bee23362489f69 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Thu, 14 Dec 2023 16:46:55 +0800
Subject: [PATCH 2/8] =?UTF-8?q?feat:=20=E6=B7=BB=E5=8A=A0gemini=E5=9F=BA?=
 =?UTF-8?q?=E7=A1=80API?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json |  2 +-
 dist/index.js       | 93 +++++++++++++++++++++++++++++++++++++++++----
 dist/timestamp      |  2 +-
 src/command.js      |  1 +
 src/context.js      | 10 ++++-
 src/env.js          | 10 ++++-
 src/gemini.js       | 67 ++++++++++++++++++++++++++++++++
 src/llm.js          |  6 +++
 8 files changed, 179 insertions(+), 12 deletions(-)
 create mode 100644 src/gemini.js

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index 30ef10b0..dabf9409 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "954408a", "timestamp": 1700550928}
+{"sha": "9a909f1", "timestamp": 1702543554}
diff --git a/dist/index.js b/dist/index.js
index 3ce11a52..e96ac80f 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -5,7 +5,7 @@ var Environment = class {
    */
   I18N = null;
   LANGUAGE = "zh-cn";
-  // AI提供商: auto, openai, azure, workers
+  // AI提供商: auto, openai, azure, workers, geminis
   AI_PROVIDER = "auto";
   // 允许访问的Telegram Token， 设置时以逗号分隔
   TELEGRAM_AVAILABLE_TOKENS = [];
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1700550928;
+  BUILD_TIMESTAMP = 1702543554;
   // 当前版本 commit id
-  BUILD_VERSION = "954408a";
+  BUILD_VERSION = "9a909f1";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -87,6 +87,12 @@ var Environment = class {
   WORKERS_CHAT_MODEL = "@cf/meta/llama-2-7b-chat-fp16";
   // Text-to-Image Model
   WORKERS_IMAGE_MODEL = "@cf/stabilityai/stable-diffusion-xl-base-1.0";
+  // Google Gemini API Key
+  GOOGLE_API_KEY = null;
+  // Google Gemini API
+  GOOGLE_COMPLETIONS_API = "https://generativelanguage.googleapis.com/v1beta/models/";
+  // Google Gemini Model
+  GOOGLE_COMPLETIONS_MODEL = "gemini.js-pro";
 };
 var ENV = new Environment();
 var DATABASE = null;
@@ -106,7 +112,8 @@ function initEnv(env, i18n2) {
     AZURE_API_KEY: "string",
     AZURE_COMPLETIONS_API: "string",
     CLOUDFLARE_ACCOUNT_ID: "string",
-    CLOUDFLARE_TOKEN: "string"
+    CLOUDFLARE_TOKEN: "string",
+    GOOGLE_API_KEY: "string"
   };
   const customCommandPrefix = "CUSTOM_COMMAND_";
   for (const key of Object.keys(env)) {
@@ -206,7 +213,13 @@ var Context = class {
     // WorkersAI聊天记录模型
     WORKERS_CHAT_MODEL: ENV.WORKERS_CHAT_MODEL,
     // WorkersAI图片模型
-    WORKER_IMAGE_MODEL: ENV.WORKERS_IMAGE_MODEL
+    WORKER_IMAGE_MODEL: ENV.WORKERS_IMAGE_MODEL,
+    // Google Gemini API Key
+    GOOGLE_API_KEY: ENV.GOOGLE_API_KEY,
+    // Google Gemini API
+    GOOGLE_COMPLETIONS_API: ENV.GOOGLE_COMPLETIONS_API,
+    // Google Gemini Model
+    GOOGLE_COMPLETIONS_MODEL: ENV.GOOGLE_COMPLETIONS_MODEL
   };
   USER_DEFINE = {
     // 自定义角色
@@ -282,7 +295,7 @@ var Context = class {
       console.error(e);
     }
     {
-      const aiProvider = new Set("auto,openai,azure,workers".split(","));
+      const aiProvider = new Set("auto,openai,azure,workers,geminis".split(","));
       if (!aiProvider.has(this.USER_CONFIG.AI_PROVIDER)) {
         this.USER_CONFIG.AI_PROVIDER = "auto";
       }
@@ -1119,7 +1132,11 @@ Body: ${JSON.stringify(body)}`);
     }
   }
   setTimeout(() => updateBotUsage(result.usage, context).catch(console.error), 0);
-  return result.choices[0].message.content;
+  try {
+    return result.choices[0].message.content;
+  } catch (e) {
+    return result?.error?.message || JSON.stringify(result);
+  }
 }
 async function requestImageFromOpenAI(prompt, context) {
   const key = openAIKeyFromContext(context);
@@ -1221,7 +1238,11 @@ ${ENV.I18N.message.loading}...`);
     return contentFull;
   } else {
     const data = await resp.json();
-    return data.result.response;
+    try {
+      return data.result.response;
+    } catch (e) {
+      return data.errors?.[0]?.message || JSON.stringify(data);
+    }
   }
 }
 async function requestImageFromWorkersAI(prompt, context) {
@@ -1229,6 +1250,56 @@ async function requestImageFromWorkersAI(prompt, context) {
   return await raw.blob();
 }
 
+// src/gemini.js
+function isGeminisAIEnable(context) {
+  return !!context.USER_CONFIG.GOOGLE_API_KEY;
+}
+async function requestCompletionsFromGeminisAI(message, history, context, onStream) {
+  const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${// 暂时不支持stream模式
+  // onStream ? 'streamGenerateContent' : 'generateContent'
+  "generateContent"}?key=${context.USER_CONFIG.GOOGLE_API_KEY}`;
+  const contentsTemp = [
+    [...history || [], { role: "user", content: message }]
+  ];
+  const contents = [];
+  for (const msg of contentsTemp) {
+    switch (msg.role) {
+      case "system":
+      case "assistant":
+        msg.role = "model";
+        break;
+      case "user":
+        msg.role = "user";
+        break;
+      default:
+        continue;
+    }
+    if (contents.length === 0 || contents[contents.length - 1].role !== msg.role) {
+      contents.push({
+        "role": msg.role,
+        "parts": [
+          {
+            "text": msg.content
+          }
+        ]
+      });
+    }
+  }
+  const resp = await fetch(url, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json"
+    },
+    body: JSON.stringify({ contents })
+  });
+  const data = await resp.json();
+  try {
+    return data.candidates[0].content.parts[0].text;
+  } catch (e) {
+    return data?.error?.message || JSON.stringify(data);
+  }
+}
+
 // src/llm.js
 async function loadHistory(key, context) {
   const initMessage = { role: "system", content: context.USER_CONFIG.SYSTEM_INIT_MESSAGE };
@@ -1301,6 +1372,8 @@ function loadChatLLM(context) {
       return requestCompletionsFromOpenAI;
     case "workers":
       return requestCompletionsFromWorkersAI;
+    case "geminis":
+      return requestCompletionsFromGeminisAI;
     default:
       if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestCompletionsFromOpenAI;
@@ -1308,6 +1381,9 @@ function loadChatLLM(context) {
       if (isWorkersAIEnable(context)) {
         return requestCompletionsFromWorkersAI;
       }
+      if (isGeminisAIEnable(context)) {
+        return requestCompletionsFromGeminisAI;
+      }
       return null;
   }
 }
@@ -1705,6 +1781,7 @@ async function commandSystem(message, command, subcommand, context) {
     context.USER_CONFIG.OPENAI_API_KEY = "******";
     context.USER_CONFIG.AZURE_API_KEY = "******";
     context.USER_CONFIG.AZURE_COMPLETIONS_API = "******";
+    context.USER_CONFIG.GOOGLE_API_KEY = "******";
     msg = "<pre>\n" + msg;
     msg += `USER_CONFIG: ${JSON.stringify(context.USER_CONFIG, null, 2)}
 `;
diff --git a/dist/timestamp b/dist/timestamp
index bac9aae7..eb94df95 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1700550928
+1702543554
diff --git a/src/command.js b/src/command.js
index 6233aa79..de111166 100644
--- a/src/command.js
+++ b/src/command.js
@@ -439,6 +439,7 @@ async function commandSystem(message, command, subcommand, context) {
     context.USER_CONFIG.OPENAI_API_KEY = '******';
     context.USER_CONFIG.AZURE_API_KEY = '******';
     context.USER_CONFIG.AZURE_COMPLETIONS_API = '******';
+    context.USER_CONFIG.GOOGLE_API_KEY = '******';
 
     msg = '<pre>\n' + msg;
     msg += `USER_CONFIG: ${JSON.stringify(context.USER_CONFIG, null, 2)}\n`;
diff --git a/src/context.js b/src/context.js
index 358db249..be05ae2b 100644
--- a/src/context.js
+++ b/src/context.js
@@ -53,6 +53,14 @@ export class Context {
     WORKERS_CHAT_MODEL: ENV.WORKERS_CHAT_MODEL,
     // WorkersAI图片模型
     WORKER_IMAGE_MODEL: ENV.WORKERS_IMAGE_MODEL,
+
+
+    // Google Gemini API Key
+    GOOGLE_API_KEY: ENV.GOOGLE_API_KEY,
+    // Google Gemini API
+    GOOGLE_COMPLETIONS_API: ENV.GOOGLE_COMPLETIONS_API,
+    // Google Gemini Model
+    GOOGLE_COMPLETIONS_MODEL: ENV.GOOGLE_COMPLETIONS_MODEL,
   };
 
   USER_DEFINE = {
@@ -118,7 +126,7 @@ export class Context {
       console.error(e);
     }
     {
-      const aiProvider = new Set('auto,openai,azure,workers'.split(','));
+      const aiProvider = new Set('auto,openai,azure,workers,geminis'.split(','));
       if (!aiProvider.has(this.USER_CONFIG.AI_PROVIDER)) {
         this.USER_CONFIG.AI_PROVIDER = 'auto';
       }
diff --git a/src/env.js b/src/env.js
index 9fa4ea68..385a596e 100644
--- a/src/env.js
+++ b/src/env.js
@@ -11,7 +11,7 @@ class Environment {
   LANGUAGE = 'zh-cn';
 
 
-  // AI提供商: auto, openai, azure, workers
+  // AI提供商: auto, openai, azure, workers, geminis
   AI_PROVIDER = 'auto';
 
   // 允许访问的Telegram Token， 设置时以逗号分隔
@@ -105,6 +105,13 @@ class Environment {
   WORKERS_CHAT_MODEL = '@cf/meta/llama-2-7b-chat-fp16';
   // Text-to-Image Model
   WORKERS_IMAGE_MODEL = '@cf/stabilityai/stable-diffusion-xl-base-1.0';
+
+  // Google Gemini API Key
+  GOOGLE_API_KEY = null;
+  // Google Gemini API
+  GOOGLE_COMPLETIONS_API = "https://generativelanguage.googleapis.com/v1beta/models/";
+  // Google Gemini Model
+  GOOGLE_COMPLETIONS_MODEL = "gemini.js-pro";
 }
 
 
@@ -139,6 +146,7 @@ export function initEnv(env, i18n) {
     AZURE_COMPLETIONS_API: 'string',
     CLOUDFLARE_ACCOUNT_ID: 'string',
     CLOUDFLARE_TOKEN: 'string',
+    GOOGLE_API_KEY: 'string',
   };
 
 
diff --git a/src/gemini.js b/src/gemini.js
new file mode 100644
index 00000000..35f6eacd
--- /dev/null
+++ b/src/gemini.js
@@ -0,0 +1,67 @@
+import {ENV} from "./env.js";
+
+
+export function isGeminisAIEnable(context) {
+    return !!(context.USER_CONFIG.GOOGLE_API_KEY);
+}
+
+/**
+ * 发送消息到Geminis
+ *
+ * @param {string} message
+ * @param {Array} history
+ * @param {Context} context
+ * @param {function} onStream
+ * @return {Promise<string>}
+ */
+export async function requestCompletionsFromGeminisAI(message, history, context, onStream) {
+    const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${
+        // 暂时不支持stream模式
+        // onStream ? 'streamGenerateContent' : 'generateContent'
+        'generateContent'
+    }?key=${context.USER_CONFIG.GOOGLE_API_KEY}`
+
+    const contentsTemp = [
+        [...history || [], {role: 'user', content: message}]
+    ]
+    const contents = []
+    // role必须是 model,user 而且不能连续两个一样
+    for (const msg of contentsTemp) {
+        switch (msg.role) {
+            case 'system':
+            case 'assistant':
+                msg.role = 'model'
+                break
+            case 'user':
+                msg.role = 'user'
+                break
+            default:
+                continue;
+        }
+        // 如果存在最后一个元素或role不一样则插入
+        if (contents.length === 0 || contents[contents.length - 1].role !== msg.role) {
+            contents.push({
+                "role": msg.role,
+                "parts": [
+                    {
+                        "text": msg.content
+                    }
+                ]
+            })
+        }
+    }
+
+    const resp = await  fetch(url, {
+        method: 'POST',
+        headers: {
+            'Content-Type': 'application/json',
+        },
+        body: JSON.stringify({contents})
+    })
+    const data = await resp.json()
+    try {
+        return data.candidates[0].content.parts[0].text
+    } catch (e) {
+        return data?.error?.message || JSON.stringify(data)
+    }
+}
diff --git a/src/llm.js b/src/llm.js
index 34622034..ea046dbe 100644
--- a/src/llm.js
+++ b/src/llm.js
@@ -9,6 +9,7 @@ import {Context} from './context.js';
 import {isAzureEnable, isOpenAIEnable, requestCompletionsFromOpenAI, requestImageFromOpenAI} from './openai.js';
 import {tokensCounter} from './utils.js';
 import {isWorkersAIEnable, requestCompletionsFromWorkersAI, requestImageFromWorkersAI} from './workers-ai.js';
+import {isGeminisAIEnable, requestCompletionsFromGeminisAI} from "./gemini.js";
 
 
 /**
@@ -117,6 +118,8 @@ export function loadChatLLM(context) {
       return requestCompletionsFromOpenAI;
     case 'workers':
       return requestCompletionsFromWorkersAI;
+    case 'geminis':
+      return requestCompletionsFromGeminisAI
     default:
       if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestCompletionsFromOpenAI;
@@ -124,6 +127,9 @@ export function loadChatLLM(context) {
       if (isWorkersAIEnable(context)) {
         return requestCompletionsFromWorkersAI;
       }
+      if (isGeminisAIEnable(context)) {
+        return requestCompletionsFromGeminisAI
+      }
       return null;
   }
 }

From 927bd09464308905cbbc9e744348a092e86a5ab6 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Thu, 14 Dec 2023 16:56:48 +0800
Subject: [PATCH 3/8] =?UTF-8?q?style:=20=E7=BB=9F=E4=B8=80gemini=E5=91=BD?=
 =?UTF-8?q?=E5=90=8D?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json |  2 +-
 dist/index.js       | 22 +++++++++++-----------
 dist/timestamp      |  2 +-
 src/context.js      |  2 +-
 src/env.js          |  2 +-
 src/gemini.js       |  6 +++---
 src/llm.js          | 10 +++++-----
 src/workers-ai.js   |  2 +-
 8 files changed, 24 insertions(+), 24 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index dabf9409..f5cb9977 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "9a909f1", "timestamp": 1702543554}
+{"sha": "6d666fd", "timestamp": 1702544157}
diff --git a/dist/index.js b/dist/index.js
index e96ac80f..a10b0e49 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -5,7 +5,7 @@ var Environment = class {
    */
   I18N = null;
   LANGUAGE = "zh-cn";
-  // AI提供商: auto, openai, azure, workers, geminis
+  // AI提供商: auto, openai, azure, workers, gemini
   AI_PROVIDER = "auto";
   // 允许访问的Telegram Token， 设置时以逗号分隔
   TELEGRAM_AVAILABLE_TOKENS = [];
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1702543554;
+  BUILD_TIMESTAMP = 1702544157;
   // 当前版本 commit id
-  BUILD_VERSION = "9a909f1";
+  BUILD_VERSION = "6d666fd";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -295,7 +295,7 @@ var Context = class {
       console.error(e);
     }
     {
-      const aiProvider = new Set("auto,openai,azure,workers,geminis".split(","));
+      const aiProvider = new Set("auto,openai,azure,workers,gemini".split(","));
       if (!aiProvider.has(this.USER_CONFIG.AI_PROVIDER)) {
         this.USER_CONFIG.AI_PROVIDER = "auto";
       }
@@ -1241,7 +1241,7 @@ ${ENV.I18N.message.loading}...`);
     try {
       return data.result.response;
     } catch (e) {
-      return data.errors?.[0]?.message || JSON.stringify(data);
+      return data?.errors?.[0]?.message || JSON.stringify(data);
     }
   }
 }
@@ -1251,10 +1251,10 @@ async function requestImageFromWorkersAI(prompt, context) {
 }
 
 // src/gemini.js
-function isGeminisAIEnable(context) {
+function isGeminiAIEnable(context) {
   return !!context.USER_CONFIG.GOOGLE_API_KEY;
 }
-async function requestCompletionsFromGeminisAI(message, history, context, onStream) {
+async function requestCompletionsFromGeminiAI(message, history, context, onStream) {
   const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${// 暂时不支持stream模式
   // onStream ? 'streamGenerateContent' : 'generateContent'
   "generateContent"}?key=${context.USER_CONFIG.GOOGLE_API_KEY}`;
@@ -1372,8 +1372,8 @@ function loadChatLLM(context) {
       return requestCompletionsFromOpenAI;
     case "workers":
       return requestCompletionsFromWorkersAI;
-    case "geminis":
-      return requestCompletionsFromGeminisAI;
+    case "gemini":
+      return requestCompletionsFromGeminiAI;
     default:
       if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestCompletionsFromOpenAI;
@@ -1381,8 +1381,8 @@ function loadChatLLM(context) {
       if (isWorkersAIEnable(context)) {
         return requestCompletionsFromWorkersAI;
       }
-      if (isGeminisAIEnable(context)) {
-        return requestCompletionsFromGeminisAI;
+      if (isGeminiAIEnable(context)) {
+        return requestCompletionsFromGeminiAI;
       }
       return null;
   }
diff --git a/dist/timestamp b/dist/timestamp
index eb94df95..81cefb79 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1702543554
+1702544157
diff --git a/src/context.js b/src/context.js
index be05ae2b..298e3297 100644
--- a/src/context.js
+++ b/src/context.js
@@ -126,7 +126,7 @@ export class Context {
       console.error(e);
     }
     {
-      const aiProvider = new Set('auto,openai,azure,workers,geminis'.split(','));
+      const aiProvider = new Set('auto,openai,azure,workers,gemini'.split(','));
       if (!aiProvider.has(this.USER_CONFIG.AI_PROVIDER)) {
         this.USER_CONFIG.AI_PROVIDER = 'auto';
       }
diff --git a/src/env.js b/src/env.js
index 385a596e..0563a443 100644
--- a/src/env.js
+++ b/src/env.js
@@ -11,7 +11,7 @@ class Environment {
   LANGUAGE = 'zh-cn';
 
 
-  // AI提供商: auto, openai, azure, workers, geminis
+  // AI提供商: auto, openai, azure, workers, gemini
   AI_PROVIDER = 'auto';
 
   // 允许访问的Telegram Token， 设置时以逗号分隔
diff --git a/src/gemini.js b/src/gemini.js
index 35f6eacd..2ee6cd9c 100644
--- a/src/gemini.js
+++ b/src/gemini.js
@@ -1,12 +1,12 @@
 import {ENV} from "./env.js";
 
 
-export function isGeminisAIEnable(context) {
+export function isGeminiAIEnable(context) {
     return !!(context.USER_CONFIG.GOOGLE_API_KEY);
 }
 
 /**
- * 发送消息到Geminis
+ * 发送消息到Gemini
  *
  * @param {string} message
  * @param {Array} history
@@ -14,7 +14,7 @@ export function isGeminisAIEnable(context) {
  * @param {function} onStream
  * @return {Promise<string>}
  */
-export async function requestCompletionsFromGeminisAI(message, history, context, onStream) {
+export async function requestCompletionsFromGeminiAI(message, history, context, onStream) {
     const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${
         // 暂时不支持stream模式
         // onStream ? 'streamGenerateContent' : 'generateContent'
diff --git a/src/llm.js b/src/llm.js
index ea046dbe..cab3de64 100644
--- a/src/llm.js
+++ b/src/llm.js
@@ -9,7 +9,7 @@ import {Context} from './context.js';
 import {isAzureEnable, isOpenAIEnable, requestCompletionsFromOpenAI, requestImageFromOpenAI} from './openai.js';
 import {tokensCounter} from './utils.js';
 import {isWorkersAIEnable, requestCompletionsFromWorkersAI, requestImageFromWorkersAI} from './workers-ai.js';
-import {isGeminisAIEnable, requestCompletionsFromGeminisAI} from "./gemini.js";
+import {isGeminiAIEnable, requestCompletionsFromGeminiAI} from "./gemini.js";
 
 
 /**
@@ -118,8 +118,8 @@ export function loadChatLLM(context) {
       return requestCompletionsFromOpenAI;
     case 'workers':
       return requestCompletionsFromWorkersAI;
-    case 'geminis':
-      return requestCompletionsFromGeminisAI
+    case 'gemini':
+      return requestCompletionsFromGeminiAI
     default:
       if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestCompletionsFromOpenAI;
@@ -127,8 +127,8 @@ export function loadChatLLM(context) {
       if (isWorkersAIEnable(context)) {
         return requestCompletionsFromWorkersAI;
       }
-      if (isGeminisAIEnable(context)) {
-        return requestCompletionsFromGeminisAI
+      if (isGeminiAIEnable(context)) {
+        return requestCompletionsFromGeminiAI
       }
       return null;
   }
diff --git a/src/workers-ai.js b/src/workers-ai.js
index 304b5d95..6a50c361 100644
--- a/src/workers-ai.js
+++ b/src/workers-ai.js
@@ -80,7 +80,7 @@ export async function requestCompletionsFromWorkersAI(message, history, context,
     try {
       return data.result.response;
     } catch (e) {
-      return data.errors?.[0]?.message || JSON.stringify(data)
+      return data?.errors?.[0]?.message || JSON.stringify(data)
     }
   }
 }

From 359c48cd872b2b35d0b9fc10130a5c16f44f1ecc Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Thu, 14 Dec 2023 18:03:17 +0800
Subject: [PATCH 4/8] =?UTF-8?q?fix:=20Gemini=E8=A7=92=E8=89=B2=E8=BD=AC?=
 =?UTF-8?q?=E6=8D=A2bug?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json | 2 +-
 dist/index.js       | 6 +++---
 dist/timestamp      | 2 +-
 src/gemini.js       | 2 +-
 4 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index f5cb9977..c3142e63 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "6d666fd", "timestamp": 1702544157}
+{"sha": "927bd09", "timestamp": 1702548165}
diff --git a/dist/index.js b/dist/index.js
index a10b0e49..4f155441 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1702544157;
+  BUILD_TIMESTAMP = 1702548165;
   // 当前版本 commit id
-  BUILD_VERSION = "6d666fd";
+  BUILD_VERSION = "927bd09";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -1264,10 +1264,10 @@ async function requestCompletionsFromGeminiAI(message, history, context, onStrea
   const contents = [];
   for (const msg of contentsTemp) {
     switch (msg.role) {
-      case "system":
       case "assistant":
         msg.role = "model";
         break;
+      case "system":
       case "user":
         msg.role = "user";
         break;
diff --git a/dist/timestamp b/dist/timestamp
index 81cefb79..8d3e3dd8 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1702544157
+1702548165
diff --git a/src/gemini.js b/src/gemini.js
index 2ee6cd9c..526cc005 100644
--- a/src/gemini.js
+++ b/src/gemini.js
@@ -28,10 +28,10 @@ export async function requestCompletionsFromGeminiAI(message, history, context,
     // role必须是 model,user 而且不能连续两个一样
     for (const msg of contentsTemp) {
         switch (msg.role) {
-            case 'system':
             case 'assistant':
                 msg.role = 'model'
                 break
+            case 'system':
             case 'user':
                 msg.role = 'user'
                 break

From b0ab11051e666e534ba39e046990503b45c7a652 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Thu, 14 Dec 2023 18:14:21 +0800
Subject: [PATCH 5/8] =?UTF-8?q?fix=EF=BC=9AGemini=E5=8E=86=E5=8F=B2?=
 =?UTF-8?q?=E8=AE=B0=E5=BD=95=E8=BD=AC=E6=8D=A2BUG?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json | 2 +-
 dist/index.js       | 8 +++-----
 dist/timestamp      | 2 +-
 src/gemini.js       | 4 +---
 4 files changed, 6 insertions(+), 10 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index c3142e63..9086ec50 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "927bd09", "timestamp": 1702548165}
+{"sha": "359c48c", "timestamp": 1702548812}
diff --git a/dist/index.js b/dist/index.js
index 4f155441..91204f8f 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1702548165;
+  BUILD_TIMESTAMP = 1702548812;
   // 当前版本 commit id
-  BUILD_VERSION = "927bd09";
+  BUILD_VERSION = "359c48c";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -1258,9 +1258,7 @@ async function requestCompletionsFromGeminiAI(message, history, context, onStrea
   const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${// 暂时不支持stream模式
   // onStream ? 'streamGenerateContent' : 'generateContent'
   "generateContent"}?key=${context.USER_CONFIG.GOOGLE_API_KEY}`;
-  const contentsTemp = [
-    [...history || [], { role: "user", content: message }]
-  ];
+  const contentsTemp = [...history || [], { role: "user", content: message }];
   const contents = [];
   for (const msg of contentsTemp) {
     switch (msg.role) {
diff --git a/dist/timestamp b/dist/timestamp
index 8d3e3dd8..0af7f7ea 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1702548165
+1702548812
diff --git a/src/gemini.js b/src/gemini.js
index 526cc005..e4e7d5c9 100644
--- a/src/gemini.js
+++ b/src/gemini.js
@@ -21,9 +21,7 @@ export async function requestCompletionsFromGeminiAI(message, history, context,
         'generateContent'
     }?key=${context.USER_CONFIG.GOOGLE_API_KEY}`
 
-    const contentsTemp = [
-        [...history || [], {role: 'user', content: message}]
-    ]
+    const contentsTemp = [...history || [], {role: 'user', content: message}]
     const contents = []
     // role必须是 model,user 而且不能连续两个一样
     for (const msg of contentsTemp) {

From 12b4944a655cf0e776edb7aa83e6071eaba528e8 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Fri, 15 Dec 2023 10:33:06 +0800
Subject: [PATCH 6/8] =?UTF-8?q?perf:=20=E4=BF=AE=E6=94=B9=E9=94=99?=
 =?UTF-8?q?=E8=AF=AF=E8=BF=94=E5=9B=9E=E9=80=BB=E8=BE=91?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json |   2 +-
 dist/index.js       |  21 ++++++---
 dist/timestamp      |   2 +-
 src/env.js          |   4 +-
 src/gemini.js       | 102 ++++++++++++++++++++++++--------------------
 src/llm.js          |   6 +--
 src/message.js      |   2 +-
 src/openai.js       |   5 ++-
 src/workers-ai.js   |   5 ++-
 9 files changed, 88 insertions(+), 61 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index 9086ec50..d6d84c25 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "359c48c", "timestamp": 1702548812}
+{"sha": "b0ab110", "timestamp": 1702607555}
diff --git a/dist/index.js b/dist/index.js
index 91204f8f..d251de86 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1702548812;
+  BUILD_TIMESTAMP = 1702607555;
   // 当前版本 commit id
-  BUILD_VERSION = "359c48c";
+  BUILD_VERSION = "b0ab110";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -1135,7 +1135,10 @@ Body: ${JSON.stringify(body)}`);
   try {
     return result.choices[0].message.content;
   } catch (e) {
-    return result?.error?.message || JSON.stringify(result);
+    if (!result) {
+      throw new Error("Empty response");
+    }
+    throw Error(result?.error?.message || JSON.stringify(result));
   }
 }
 async function requestImageFromOpenAI(prompt, context) {
@@ -1241,7 +1244,10 @@ ${ENV.I18N.message.loading}...`);
     try {
       return data.result.response;
     } catch (e) {
-      return data?.errors?.[0]?.message || JSON.stringify(data);
+      if (!data) {
+        throw new Error("Empty response");
+      }
+      throw new Error(data?.errors?.[0]?.message || JSON.stringify(data));
     }
   }
 }
@@ -1281,6 +1287,8 @@ async function requestCompletionsFromGeminiAI(message, history, context, onStrea
           }
         ]
       });
+    } else {
+      contents[contents.length - 1].parts[0].text += msg.content;
     }
   }
   const resp = await fetch(url, {
@@ -1294,7 +1302,10 @@ async function requestCompletionsFromGeminiAI(message, history, context, onStrea
   try {
     return data.candidates[0].content.parts[0].text;
   } catch (e) {
-    return data?.error?.message || JSON.stringify(data);
+    if (!data) {
+      throw new Error("Empty response");
+    }
+    throw new Error(data?.error?.message || JSON.stringify(data));
   }
 }
 
diff --git a/dist/timestamp b/dist/timestamp
index 0af7f7ea..85b8d56f 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1702548812
+1702607555
diff --git a/src/env.js b/src/env.js
index 0563a443..4c3e457b 100644
--- a/src/env.js
+++ b/src/env.js
@@ -109,9 +109,9 @@ class Environment {
   // Google Gemini API Key
   GOOGLE_API_KEY = null;
   // Google Gemini API
-  GOOGLE_COMPLETIONS_API = "https://generativelanguage.googleapis.com/v1beta/models/";
+  GOOGLE_COMPLETIONS_API = 'https://generativelanguage.googleapis.com/v1beta/models/';
   // Google Gemini Model
-  GOOGLE_COMPLETIONS_MODEL = "gemini.js-pro";
+  GOOGLE_COMPLETIONS_MODEL = 'gemini.js-pro';
 }
 
 
diff --git a/src/gemini.js b/src/gemini.js
index e4e7d5c9..2a83de21 100644
--- a/src/gemini.js
+++ b/src/gemini.js
@@ -1,8 +1,12 @@
-import {ENV} from "./env.js";
-
+/* eslint-disable no-unused-vars */
+import {Context} from './context.js';
 
+/**
+ * @param {Context} context
+ * @return {boolean}
+ */
 export function isGeminiAIEnable(context) {
-    return !!(context.USER_CONFIG.GOOGLE_API_KEY);
+  return !!(context.USER_CONFIG.GOOGLE_API_KEY);
 }
 
 /**
@@ -15,51 +19,57 @@ export function isGeminiAIEnable(context) {
  * @return {Promise<string>}
  */
 export async function requestCompletionsFromGeminiAI(message, history, context, onStream) {
-    const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${
-        // 暂时不支持stream模式
-        // onStream ? 'streamGenerateContent' : 'generateContent'
-        'generateContent'
-    }?key=${context.USER_CONFIG.GOOGLE_API_KEY}`
+  const url = `${context.USER_CONFIG.GOOGLE_COMPLETIONS_API}${context.USER_CONFIG.GOOGLE_COMPLETIONS_MODEL}:${
+    // 暂时不支持stream模式
+    // onStream ? 'streamGenerateContent' : 'generateContent'
+    'generateContent'
+  }?key=${context.USER_CONFIG.GOOGLE_API_KEY}`;
 
-    const contentsTemp = [...history || [], {role: 'user', content: message}]
-    const contents = []
-    // role必须是 model,user 而且不能连续两个一样
-    for (const msg of contentsTemp) {
-        switch (msg.role) {
-            case 'assistant':
-                msg.role = 'model'
-                break
-            case 'system':
-            case 'user':
-                msg.role = 'user'
-                break
-            default:
-                continue;
-        }
-        // 如果存在最后一个元素或role不一样则插入
-        if (contents.length === 0 || contents[contents.length - 1].role !== msg.role) {
-            contents.push({
-                "role": msg.role,
-                "parts": [
-                    {
-                        "text": msg.content
-                    }
-                ]
-            })
-        }
+  const contentsTemp = [...history || [], {role: 'user', content: message}];
+  const contents = [];
+  // role必须是 model,user 而且不能连续两个一样
+  for (const msg of contentsTemp) {
+    switch (msg.role) {
+      case 'assistant':
+        msg.role = 'model';
+        break;
+      case 'system':
+      case 'user':
+        msg.role = 'user';
+        break;
+      default:
+        continue;
+    }
+    // 如果存在最后一个元素或role不一样则插入
+    if (contents.length === 0 || contents[contents.length - 1].role !== msg.role) {
+      contents.push({
+        'role': msg.role,
+        'parts': [
+          {
+            'text': msg.content,
+          },
+        ],
+      });
+    } else {
+      // 否则合并
+      contents[contents.length - 1].parts[0].text += msg.content;
     }
+  }
 
-    const resp = await  fetch(url, {
-        method: 'POST',
-        headers: {
-            'Content-Type': 'application/json',
-        },
-        body: JSON.stringify({contents})
-    })
-    const data = await resp.json()
-    try {
-        return data.candidates[0].content.parts[0].text
-    } catch (e) {
-        return data?.error?.message || JSON.stringify(data)
+  const resp = await fetch(url, {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+    },
+    body: JSON.stringify({contents}),
+  });
+  const data = await resp.json();
+  try {
+    return data.candidates[0].content.parts[0].text;
+  } catch (e) {
+    if (!data) {
+      throw new Error('Empty response');
     }
+    throw new Error(data?.error?.message || JSON.stringify(data));
+  }
 }
diff --git a/src/llm.js b/src/llm.js
index cab3de64..b05c5e41 100644
--- a/src/llm.js
+++ b/src/llm.js
@@ -9,7 +9,7 @@ import {Context} from './context.js';
 import {isAzureEnable, isOpenAIEnable, requestCompletionsFromOpenAI, requestImageFromOpenAI} from './openai.js';
 import {tokensCounter} from './utils.js';
 import {isWorkersAIEnable, requestCompletionsFromWorkersAI, requestImageFromWorkersAI} from './workers-ai.js';
-import {isGeminiAIEnable, requestCompletionsFromGeminiAI} from "./gemini.js";
+import {isGeminiAIEnable, requestCompletionsFromGeminiAI} from './gemini.js';
 
 
 /**
@@ -119,7 +119,7 @@ export function loadChatLLM(context) {
     case 'workers':
       return requestCompletionsFromWorkersAI;
     case 'gemini':
-      return requestCompletionsFromGeminiAI
+      return requestCompletionsFromGeminiAI;
     default:
       if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestCompletionsFromOpenAI;
@@ -128,7 +128,7 @@ export function loadChatLLM(context) {
         return requestCompletionsFromWorkersAI;
       }
       if (isGeminiAIEnable(context)) {
-        return requestCompletionsFromGeminiAI
+        return requestCompletionsFromGeminiAI;
       }
       return null;
   }
diff --git a/src/message.js b/src/message.js
index a9e1c42a..da12478c 100644
--- a/src/message.js
+++ b/src/message.js
@@ -257,7 +257,7 @@ async function msgHandleRole(message, context) {
     for (const key in roleConfig) {
       if ( context.USER_CONFIG.hasOwnProperty(key) && typeof context.USER_CONFIG[key] === typeof roleConfig[key] ) {
         if (ENV.LOCK_USER_CONFIG_KEYS.includes(key)) {
-            continue;
+          continue;
         }
         context.USER_CONFIG[key] = roleConfig[key];
       }
diff --git a/src/openai.js b/src/openai.js
index 9feb5740..89088383 100644
--- a/src/openai.js
+++ b/src/openai.js
@@ -123,7 +123,10 @@ export async function requestCompletionsFromOpenAI(message, history, context, on
   try {
     return result.choices[0].message.content;
   } catch (e) {
-    return  result?.error?.message || JSON.stringify(result);
+    if (!result) {
+      throw new Error('Empty response');
+    }
+    throw Error(result?.error?.message || JSON.stringify(result));
   }
 }
 
diff --git a/src/workers-ai.js b/src/workers-ai.js
index 6a50c361..93c6e632 100644
--- a/src/workers-ai.js
+++ b/src/workers-ai.js
@@ -80,7 +80,10 @@ export async function requestCompletionsFromWorkersAI(message, history, context,
     try {
       return data.result.response;
     } catch (e) {
-      return data?.errors?.[0]?.message || JSON.stringify(data)
+      if (!data) {
+        throw new Error('Empty response');
+      }
+      throw new Error(data?.errors?.[0]?.message || JSON.stringify(data));
     }
   }
 }

From 9273c897e5e002d5d45d3d057c5d8f230aadbdb5 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Fri, 12 Jan 2024 11:07:10 +0800
Subject: [PATCH 7/8] =?UTF-8?q?feat:=20=E6=B7=BB=E5=8A=A0=20AZURE=5FDALLE?=
 =?UTF-8?q?=5FAPI=20=E6=94=AF=E6=8C=81?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json |  2 +-
 dist/index.js       | 50 +++++++++++++++++++++++++++++++--------------
 dist/timestamp      |  2 +-
 src/command.js      |  1 +
 src/context.js      |  2 ++
 src/env.js          |  3 +++
 src/gemini.js       |  2 ++
 src/llm.js          |  4 ++--
 src/openai.js       | 35 ++++++++++++++++++++++---------
 9 files changed, 72 insertions(+), 29 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index d6d84c25..82a5b0d8 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "b0ab110", "timestamp": 1702607555}
+{"sha": "12b4944", "timestamp": 1705028785}
diff --git a/dist/index.js b/dist/index.js
index d251de86..30629fa5 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -58,9 +58,9 @@ var Environment = class {
   // 检查更新的分支
   UPDATE_BRANCH = "master";
   // 当前版本
-  BUILD_TIMESTAMP = 1702607555;
+  BUILD_TIMESTAMP = 1705028785;
   // 当前版本 commit id
-  BUILD_VERSION = "b0ab110";
+  BUILD_VERSION = "12b4944";
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -79,6 +79,8 @@ var Environment = class {
   AZURE_API_KEY = null;
   // Azure Completions API
   AZURE_COMPLETIONS_API = null;
+  // Azure DallE API
+  AZURE_DALLE_API = null;
   // Cloudflare Account ID
   CLOUDFLARE_ACCOUNT_ID = null;
   // Cloudflare Token
@@ -111,6 +113,7 @@ function initEnv(env, i18n2) {
     OPENAI_API_BASE: "string",
     AZURE_API_KEY: "string",
     AZURE_COMPLETIONS_API: "string",
+    AZURE_DALLE_API: "string",
     CLOUDFLARE_ACCOUNT_ID: "string",
     CLOUDFLARE_TOKEN: "string",
     GOOGLE_API_KEY: "string"
@@ -210,6 +213,8 @@ var Context = class {
     AZURE_API_KEY: ENV.AZURE_API_KEY,
     // Azure Completions API
     AZURE_COMPLETIONS_API: ENV.AZURE_COMPLETIONS_API,
+    // Azure DALL-E API
+    AZURE_DALLE_API: ENV.AZURE_DALLE_API,
     // WorkersAI聊天记录模型
     WORKERS_CHAT_MODEL: ENV.WORKERS_CHAT_MODEL,
     // WorkersAI图片模型
@@ -1062,9 +1067,8 @@ function isOpenAIEnable(context) {
   return context.USER_CONFIG.OPENAI_API_KEY || ENV.API_KEY.length > 0;
 }
 function isAzureEnable(context) {
-  const api = context.USER_CONFIG.AZURE_COMPLETIONS_API || ENV.AZURE_COMPLETIONS_API;
   const key = context.USER_CONFIG.AZURE_API_KEY || ENV.AZURE_API_KEY;
-  return api !== null && key !== null;
+  return key !== null;
 }
 async function requestCompletionsFromOpenAI(message, history, context, onStream) {
   const body = {
@@ -1084,8 +1088,8 @@ async function requestCompletionsFromOpenAI(message, history, context, onStream)
   };
   {
     const provider = context.USER_CONFIG.AI_PROVIDER;
-    if (provider === "azure" || provider === "auto" && isAzureEnable(context)) {
-      url = ENV.AZURE_COMPLETIONS_API;
+    if (provider === "azure" || provider === "auto" && isAzureEnable(context) && context.USER_CONFIG.AZURE_COMPLETIONS_API !== null) {
+      url = context.USER_CONFIG.AZURE_COMPLETIONS_API;
       header["api-key"] = azureKeyFromContext(context);
       delete header["Authorization"];
       delete body.model;
@@ -1142,7 +1146,11 @@ Body: ${JSON.stringify(body)}`);
   }
 }
 async function requestImageFromOpenAI(prompt, context) {
-  const key = openAIKeyFromContext(context);
+  let url = `${ENV.OPENAI_API_BASE}/images/generations`;
+  const header = {
+    "Content-Type": "application/json",
+    "Authorization": `Bearer ${openAIKeyFromContext(context)}`
+  };
   const body = {
     prompt,
     n: 1,
@@ -1153,12 +1161,22 @@ async function requestImageFromOpenAI(prompt, context) {
     body.quality = context.USER_CONFIG.DALL_E_IMAGE_QUALITY;
     body.style = context.USER_CONFIG.DALL_E_IMAGE_STYLE;
   }
-  const resp = await fetch(`${ENV.OPENAI_API_BASE}/images/generations`, {
+  {
+    const provider = context.USER_CONFIG.AI_PROVIDER;
+    if (provider === "azure" || provider === "auto" && isAzureEnable(context) && context.USER_CONFIG.AZURE_DALLE_API !== null) {
+      url = context.USER_CONFIG.AZURE_DALLE_API;
+      const vaildSize = ["1792x1024", "1024x1024", "1024x1792"];
+      if (!vaildSize.includes(body.size)) {
+        body.size = "1024x1024";
+      }
+      header["api-key"] = azureKeyFromContext(context);
+      delete header["Authorization"];
+      delete body.model;
+    }
+  }
+  const resp = await fetch(url, {
     method: "POST",
-    headers: {
-      "Content-Type": "application/json",
-      "Authorization": `Bearer ${key}`
-    },
+    headers: header,
     body: JSON.stringify(body)
   }).then((res) => res.json());
   if (resp.error?.message) {
@@ -1294,7 +1312,8 @@ async function requestCompletionsFromGeminiAI(message, history, context, onStrea
   const resp = await fetch(url, {
     method: "POST",
     headers: {
-      "Content-Type": "application/json"
+      "Content-Type": "application/json",
+      "User-Agent": CONST.USER_AGENT
     },
     body: JSON.stringify({ contents })
   });
@@ -1401,11 +1420,11 @@ function loadImageGen(context) {
     case "openai":
       return requestImageFromOpenAI;
     case "azure":
-      return null;
+      return requestImageFromOpenAI;
     case "workers":
       return requestImageFromWorkersAI;
     default:
-      if (isOpenAIEnable(context)) {
+      if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestImageFromOpenAI;
       }
       if (isWorkersAIEnable(context)) {
@@ -1790,6 +1809,7 @@ async function commandSystem(message, command, subcommand, context) {
     context.USER_CONFIG.OPENAI_API_KEY = "******";
     context.USER_CONFIG.AZURE_API_KEY = "******";
     context.USER_CONFIG.AZURE_COMPLETIONS_API = "******";
+    context.USER_CONFIG.AZURE_DALLE_API = "******";
     context.USER_CONFIG.GOOGLE_API_KEY = "******";
     msg = "<pre>\n" + msg;
     msg += `USER_CONFIG: ${JSON.stringify(context.USER_CONFIG, null, 2)}
diff --git a/dist/timestamp b/dist/timestamp
index 85b8d56f..52b2a207 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1702607555
+1705028785
diff --git a/src/command.js b/src/command.js
index de111166..f06c4c8d 100644
--- a/src/command.js
+++ b/src/command.js
@@ -439,6 +439,7 @@ async function commandSystem(message, command, subcommand, context) {
     context.USER_CONFIG.OPENAI_API_KEY = '******';
     context.USER_CONFIG.AZURE_API_KEY = '******';
     context.USER_CONFIG.AZURE_COMPLETIONS_API = '******';
+    context.USER_CONFIG.AZURE_DALLE_API = '******';
     context.USER_CONFIG.GOOGLE_API_KEY = '******';
 
     msg = '<pre>\n' + msg;
diff --git a/src/context.js b/src/context.js
index 298e3297..328e1417 100644
--- a/src/context.js
+++ b/src/context.js
@@ -48,6 +48,8 @@ export class Context {
     AZURE_API_KEY: ENV.AZURE_API_KEY,
     // Azure Completions API
     AZURE_COMPLETIONS_API: ENV.AZURE_COMPLETIONS_API,
+    // Azure DALL-E API
+    AZURE_DALLE_API: ENV.AZURE_DALLE_API,
 
     // WorkersAI聊天记录模型
     WORKERS_CHAT_MODEL: ENV.WORKERS_CHAT_MODEL,
diff --git a/src/env.js b/src/env.js
index 4c3e457b..a5a61c53 100644
--- a/src/env.js
+++ b/src/env.js
@@ -96,6 +96,8 @@ class Environment {
   AZURE_API_KEY = null;
   // Azure Completions API
   AZURE_COMPLETIONS_API = null;
+  // Azure DallE API
+  AZURE_DALLE_API = null;
 
   // Cloudflare Account ID
   CLOUDFLARE_ACCOUNT_ID = null;
@@ -144,6 +146,7 @@ export function initEnv(env, i18n) {
     OPENAI_API_BASE: 'string',
     AZURE_API_KEY: 'string',
     AZURE_COMPLETIONS_API: 'string',
+    AZURE_DALLE_API: 'string',
     CLOUDFLARE_ACCOUNT_ID: 'string',
     CLOUDFLARE_TOKEN: 'string',
     GOOGLE_API_KEY: 'string',
diff --git a/src/gemini.js b/src/gemini.js
index 2a83de21..6f2e57ef 100644
--- a/src/gemini.js
+++ b/src/gemini.js
@@ -1,5 +1,6 @@
 /* eslint-disable no-unused-vars */
 import {Context} from './context.js';
+import {CONST} from './env.js';
 
 /**
  * @param {Context} context
@@ -60,6 +61,7 @@ export async function requestCompletionsFromGeminiAI(message, history, context,
     method: 'POST',
     headers: {
       'Content-Type': 'application/json',
+      'User-Agent': CONST.USER_AGENT,
     },
     body: JSON.stringify({contents}),
   });
diff --git a/src/llm.js b/src/llm.js
index b05c5e41..c7543dd2 100644
--- a/src/llm.js
+++ b/src/llm.js
@@ -144,11 +144,11 @@ export function loadImageGen(context) {
     case 'openai':
       return requestImageFromOpenAI;
     case 'azure':
-      return null;
+      return requestImageFromOpenAI;
     case 'workers':
       return requestImageFromWorkersAI;
     default:
-      if (isOpenAIEnable(context)) {
+      if (isOpenAIEnable(context) || isAzureEnable(context)) {
         return requestImageFromOpenAI;
       }
       if (isWorkersAIEnable(context)) {
diff --git a/src/openai.js b/src/openai.js
index 89088383..aafc0ab0 100644
--- a/src/openai.js
+++ b/src/openai.js
@@ -40,9 +40,9 @@ export function isOpenAIEnable(context) {
  * @return {boolean}
  */
 export function isAzureEnable(context) {
-  const api = context.USER_CONFIG.AZURE_COMPLETIONS_API || ENV.AZURE_COMPLETIONS_API;
+  // const api = context.USER_CONFIG.AZURE_COMPLETIONS_API || ENV.AZURE_COMPLETIONS_API;
   const key = context.USER_CONFIG.AZURE_API_KEY || ENV.AZURE_API_KEY;
-  return api !== null && key !== null;
+  return key !== null;
 }
 
 
@@ -75,8 +75,8 @@ export async function requestCompletionsFromOpenAI(message, history, context, on
   };
   {
     const provider = context.USER_CONFIG.AI_PROVIDER;
-    if (provider === 'azure' || (provider === 'auto' && isAzureEnable(context)) ) {
-      url = ENV.AZURE_COMPLETIONS_API;
+    if (provider === 'azure' || (provider === 'auto' && isAzureEnable(context) && context.USER_CONFIG.AZURE_COMPLETIONS_API !== null) ) {
+      url = context.USER_CONFIG.AZURE_COMPLETIONS_API;
       header['api-key'] = azureKeyFromContext(context);
       delete header['Authorization'];
       delete body.model;
@@ -138,7 +138,11 @@ export async function requestCompletionsFromOpenAI(message, history, context, on
  * @return {Promise<string>}
  */
 export async function requestImageFromOpenAI(prompt, context) {
-  const key = openAIKeyFromContext(context);
+  let url = `${ENV.OPENAI_API_BASE}/images/generations`;
+  const header = {
+    'Content-Type': 'application/json',
+    'Authorization': `Bearer ${openAIKeyFromContext(context)}`,
+  };
   const body = {
     prompt: prompt,
     n: 1,
@@ -149,12 +153,23 @@ export async function requestImageFromOpenAI(prompt, context) {
     body.quality = context.USER_CONFIG.DALL_E_IMAGE_QUALITY;
     body.style = context.USER_CONFIG.DALL_E_IMAGE_STYLE;
   }
-  const resp = await fetch(`${ENV.OPENAI_API_BASE}/images/generations`, {
+  {
+    const provider = context.USER_CONFIG.AI_PROVIDER;
+    if (provider === 'azure' || (provider === 'auto' && isAzureEnable(context) && context.USER_CONFIG.AZURE_DALLE_API !== null) ) {
+      url = context.USER_CONFIG.AZURE_DALLE_API;
+      // 1792x1024, 1024x1024, or 1024x1792.
+      const vaildSize = ['1792x1024', '1024x1024', '1024x1792'];
+      if (!vaildSize.includes(body.size)) {
+        body.size = '1024x1024';
+      }
+      header['api-key'] = azureKeyFromContext(context);
+      delete header['Authorization'];
+      delete body.model;
+    }
+  }
+  const resp = await fetch(url, {
     method: 'POST',
-    headers: {
-      'Content-Type': 'application/json',
-      'Authorization': `Bearer ${key}`,
-    },
+    headers: header,
     body: JSON.stringify(body),
   }).then((res) => res.json());
   if (resp.error?.message) {

From 8c71bddabb8f44c6fd0e37eb40bd4d787c796225 Mon Sep 17 00:00:00 2001
From: TBXark <tbxark@outlook.com>
Date: Mon, 15 Jan 2024 14:40:28 +0800
Subject: [PATCH 8/8] =?UTF-8?q?doc:=20=E4=BC=98=E5=8C=96=E9=85=8D=E7=BD=AE?=
 =?UTF-8?q?=E6=96=87=E6=A1=A3?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 dist/buildinfo.json |   2 +-
 dist/index.js       |  58 +++++++++++++------
 dist/timestamp      |   2 +-
 doc/cn/CONFIG.md    | 132 ++++++++++++++++++++++---------------------
 doc/en/CONFIG.md    | 133 ++++++++++++++++++++++----------------------
 src/env.js          |  67 ++++++++++++++--------
 6 files changed, 224 insertions(+), 170 deletions(-)

diff --git a/dist/buildinfo.json b/dist/buildinfo.json
index 82a5b0d8..a7e47d52 100644
--- a/dist/buildinfo.json
+++ b/dist/buildinfo.json
@@ -1 +1 @@
-{"sha": "12b4944", "timestamp": 1705028785}
+{"sha": "9273c89", "timestamp": 1705300786}
diff --git a/dist/index.js b/dist/index.js
index 30629fa5..82aeb624 100644
--- a/dist/index.js
+++ b/dist/index.js
@@ -1,20 +1,38 @@
 // src/env.js
 var Environment = class {
+  // -- 版本数据 --
+  //
+  // 当前版本
+  BUILD_TIMESTAMP = 1705300786;
+  // 当前版本 commit id
+  BUILD_VERSION = "9273c89";
+  // -- 基础配置 --
   /**
    * @type {I18n | null}
    */
   I18N = null;
+  // 多语言支持
   LANGUAGE = "zh-cn";
+  // 检查更新的分支
+  UPDATE_BRANCH = "master";
   // AI提供商: auto, openai, azure, workers, gemini
   AI_PROVIDER = "auto";
+  // -- Telegram 相关 --
+  //
+  // Telegram API Domain
+  TELEGRAM_API_DOMAIN = "https://api.telegram.org";
   // 允许访问的Telegram Token， 设置时以逗号分隔
   TELEGRAM_AVAILABLE_TOKENS = [];
+  // --  权限相关 --
+  //
   // 允许所有人使用
   I_AM_A_GENEROUS_PERSON = false;
   // 白名单
   CHAT_WHITE_LIST = [];
   // 用户配置
   LOCK_USER_CONFIG_KEYS = [];
+  // -- 群组相关 --
+  //
   // 允许访问的Telegram Token 对应的Bot Name， 设置时以逗号分隔
   TELEGRAM_BOT_NAME = [];
   // 群组白名单
@@ -23,10 +41,8 @@ var Environment = class {
   GROUP_CHAT_BOT_ENABLE = true;
   // 群组机器人共享模式,关闭后，一个群组只有一个会话和配置。开启的话群组的每个人都有自己的会话上下文
   GROUP_CHAT_BOT_SHARE_MODE = false;
-  // OpenAI API Key
-  API_KEY = [];
-  // OpenAI的模型名称
-  CHAT_MODEL = "gpt-3.5-turbo";
+  // -- 历史记录相关 --
+  //
   // 为了避免4096字符限制，将消息删减
   AUTO_TRIM_HISTORY = true;
   // 最大历史记录长度
@@ -37,10 +53,24 @@ var Environment = class {
   GPT3_TOKENS_COUNT = false;
   // GPT3计数器资源地址
   GPT3_TOKENS_COUNT_REPO = "https://raw.githubusercontent.com/tbxark-arc/GPT-3-Encoder/master";
+  // -- Prompt 相关 --
+  //
   // 全局默认初始化消息
   SYSTEM_INIT_MESSAGE = null;
   // 全局默认初始化消息角色
   SYSTEM_INIT_MESSAGE_ROLE = "system";
+  // -- Open AI 配置 --
+  //
+  // OpenAI API Key
+  API_KEY = [];
+  // OpenAI的模型名称
+  CHAT_MODEL = "gpt-3.5-turbo";
+  // OpenAI API Domain 可替换兼容openai api的其他服务商
+  OPENAI_API_DOMAIN = "https://api.openai.com";
+  // OpenAI API BASE `https://api.openai.com/v1`
+  OPENAI_API_BASE = null;
+  // -- DALLE 配置 --
+  //
   // DALL-E的模型名称
   DALL_E_MODEL = "dall-e-2";
   // DALL-E图片尺寸
@@ -49,18 +79,16 @@ var Environment = class {
   DALL_E_IMAGE_QUALITY = "standard";
   // DALL-E图片风格
   DALL_E_IMAGE_STYLE = "vivid";
+  // -- 特性开关 --
+  //
   // 是否开启使用统计
   ENABLE_USAGE_STATISTICS = false;
   // 隐藏部分命令按钮
   HIDE_COMMAND_BUTTONS = ["/role"];
   // 显示快捷回复按钮
   SHOW_REPLY_BUTTON = false;
-  // 检查更新的分支
-  UPDATE_BRANCH = "master";
-  // 当前版本
-  BUILD_TIMESTAMP = 1705028785;
-  // 当前版本 commit id
-  BUILD_VERSION = "12b4944";
+  // -- 模式开关 --
+  //
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -69,12 +97,8 @@ var Environment = class {
   DEBUG_MODE = false;
   // 开发模式
   DEV_MODE = false;
-  // Telegram API Domain
-  TELEGRAM_API_DOMAIN = "https://api.telegram.org";
-  // OpenAI API Domain 可替换兼容openai api的其他服务商
-  OPENAI_API_DOMAIN = "https://api.openai.com";
-  // OpenAI API BASE `https://api.openai.com/v1`
-  OPENAI_API_BASE = null;
+  // -- AZURE 配置 --
+  //
   // Azure API Key
   AZURE_API_KEY = null;
   // Azure Completions API
@@ -86,7 +110,7 @@ var Environment = class {
   // Cloudflare Token
   CLOUDFLARE_TOKEN = null;
   // Text Generation Model
-  WORKERS_CHAT_MODEL = "@cf/meta/llama-2-7b-chat-fp16";
+  WORKERS_CHAT_MODEL = "@cf/mistral/mistral-7b-instruct-v0.1 ";
   // Text-to-Image Model
   WORKERS_IMAGE_MODEL = "@cf/stabilityai/stable-diffusion-xl-base-1.0";
   // Google Gemini API Key
diff --git a/dist/timestamp b/dist/timestamp
index 52b2a207..6e157c85 100644
--- a/dist/timestamp
+++ b/dist/timestamp
@@ -1 +1 @@
-1705028785
+1705300786
diff --git a/doc/cn/CONFIG.md b/doc/cn/CONFIG.md
index d47f8212..a76c3845 100644
--- a/doc/cn/CONFIG.md
+++ b/doc/cn/CONFIG.md
@@ -14,54 +14,55 @@
 
 为每个用户通用的配置，通常在workers配置界面填写
 
-| KEY                       | 说明                     | 默认值                                            | 特殊说明                                                                                                                            |
-|:--------------------------|------------------------|------------------------------------------------|---------------------------------------------------------------------------------------------------------------------------------|
-| AI_PROVIDER               | AI提供商                  | `auto`                                         | AI提供商: auto, azure, openai, workers; auto为自动选择一个有效的配置，判断优先级为 azure > openai > workers                                           |
-| API_KEY                   | OpenAI API Key         | `null`                                         | 可以同时使用多个key，使用的时候会随机选择一个                                                                                                        |
-| CHAT_MODEL                | open ai 模型选择           | `gpt-3.5-turbo`                                |                                                                                                                                 |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| TELEGRAM_AVAILABLE_TOKENS | 支持多个Telegram Bot Token | `null`                                         | 多个Token用`,`分隔                                                                                                                   |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| CHAT_WHITE_LIST           | 聊天ID白名单                | `null`                                         | 多个ID用`,`分隔，不知道ID，和机器人聊一句就能返回                                                                                                    |
-| I_AM_A_GENEROUS_PERSON    | 关闭白名单，允许所有人访问          | `false`                                        | 鉴于很多人不想设置白名单，或者不知道怎么获取ID，所以设置这个选项就能允许所有人访问， 值为`true`时生效                                                                         |
-| LOCK_USER_CONFIG_KEYS     | 锁定自定义用户配置              | `[]`                                           | 可以锁定某些字段。比如设置为`CHAT_MODEL`就可以防止其他用户通过`/setenv`指令切换模型，多个字段用`,`分隔                                                                 |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| AUTO_TRIM_HISTORY         | 自动清理历史记录               | `true`                                         | 为了避免4096字符限制，将消息删减                                                                                                              |
-| MAX_HISTORY_LENGTH        | 最大历史记录长度               | `20`                                           | `AUTO_TRIM_HISTORY开启后` 为了避免4096字符限制，将消息删减                                                                                       |
-| MAX_TOKEN_LENGTH          | 最大历史token数量            | 2048                                           | 过长容易超时建议设定在一个合适的数字                                                                                                              |
-| GPT3_TOKENS_COUNT         | GTP计数模式                | `false`                                        | 使用更加精准的token计数模式替代单纯判断字符串长度，但是容易超时                                                                                              |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| SYSTEM_INIT_MESSAGE       | 系统初始化信息                | `你是一个得力的助手`                                    | 默认机器人设定                                                                                                                         |
-| SYSTEM_INIT_MESSAGE_ROLE  | 系统初始化信息角色              | `system`                                       | 默认机器人设定                                                                                                                         |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| ENABLE_USAGE_STATISTICS   | 开启使用统计                 | `false`                                        | 开启后，每次调用API都会记录到KV，可以通过`/usage`查看                                                                                               |
-| HIDE_COMMAND_BUTTONS      | 隐藏指令按钮                 | `null`                                         | 把想要隐藏的按钮写入用逗号分开`/start,/system`, 记得带上斜杠,修改之后得重新`init`                                                                           |
-| SHOW_REPLY_BUTTON         | 显示快捷回复按钮               | `false`                                        | 显示快捷回复按钮                                                                                                                        |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| UPDATE_BRANCH             | 分支                     | `master`                                       | 版本检测所在分支                                                                                                                        |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| DEBUG_MODE                | 调试模式                   | `false`                                        | 目前可以把最新一条消息保存到KV方便调试,非常消耗KV写入量，正式环境务必关闭                                                                                         |
-| DEV_MODE                  | 开发模式                   | `false`                                        | 开发测试用                                                                                                                           |
-| STREAM_MODE               | 流模式                    | `true`                                         | 得到类似ChatGPT Web一样的打字机输出模式                                                                                                       |
-| SAFE_MODE                 | 安全模式                   | `true`                                         | 安全模式，会增加KV写损耗，但是能避免Workers超时导致的Telegram死亡循环重试，减少Token的浪费，不建议关闭。                                                                 |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| LANGUAGE                  | 语言                     | `zh-CN`                                        | `zh-CN`，`zh-TW`和`en`                                                                                                            |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| TELEGRAM_API_DOMAIN       | Telegram               | `https://api.telegram.org`                     | 可以自定义Telegram服务器                                                                                                                |
-| OPENAI_API_DOMAIN         | OpenAI                 | `https://api.openai.com`                       | 可以替换为其他与OpenAI API兼容的其他服务商的域名                                                                                                   |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| AZURE_API_KEY             | azure api key          | `null`                                         | 支持azure的API，两个密钥随便选一个就可以。如果你要默认使用azure，你可以设置`AI_PROVIDER`为`azure`                                                               |
-| AZURE_COMPLETIONS_API     | azure api url          | `null`                                         | 格式`https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/chat/completions?api-version=2023-05-15` |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| CLOUDFLARE_ACCOUNT_ID     | Cloudflare 的 用户ID      | `null`                                         | 你可以在workers首页的右侧信息栏中找到这个信息。如果你要默认使用workers ai，你可以设置`AI_PROVIDER`为`workers`                                                      |
-| CLOUDFLARE_TOKEN          | Cloudflare的Token       | `null`                                         | 你可以在`https://dash.cloudflare.com/profile/api-tokens`中使用`Workers AI (Beta)`模板创建                                                  |
-| WORKERS_CHAT_MODEL        | 文字生成模型                 | `@cf/meta/llama-2-7b-chat-fp16`                | 具体模型列表可以查看`https://developers.cloudflare.com/workers-ai/models/llm/`                                                            |
-| WORKERS_IMAGE_MODEL       | 文字生成图片模型               | `@cf/stabilityai/stable-diffusion-xl-base-1.0` | 同上                                                                                                                              |
-| -                         | -                      | -                                              | -                                                                                                                               |
-| DALL_E_MODEL              | 生成图像的模型                | `dall-e-2`                                     | 支持 `dall-e-2` 和 `dall-e-3`                                                                                                      |
-| DALL_E_IMAGE_SIZE         | 生成图像的尺寸                | `512x512`                                      | 生成图像的尺寸。对于 dall-e-2，必须是256x256、512x512或1024x1024之一。对于 dall-e-3 模型，必须是1024x1024、1792x1024或1024x1792之一。                           |
-| DALL_E_IMAGE_QUALITY      | 生成图像的质量                | `standard`                                     | 将要生成的图片质量。hd会创建具有更精细细节和整体一致性的图片。此参数仅支持dall-e-3.                                                                                 |
-| DALL_E_IMAGE_STYLE        | 生成图像的风格                | `vivid`                                        | 生成图像的风格。必须是 vivid 或 natural 中的一个。vivid使模型倾向于产生超现实和戏剧化的图片。natural使模型产生更自然、不那么超现实外观的图片。此参数仅支持dall-e-3.                            |
+| KEY                       | 说明                                         | 默认值                                                                 | 特殊说明                                                                                                                            |
+|:--------------------------|--------------------------------------------|---------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------------------------|
+| LANGUAGE                  | 语言                                         | `zh-CN`                                                             | `zh-CN`，`zh-TW`和`en`                                                                                                            |
+| AI_PROVIDER               | AI提供商                                      | `auto`                                                              | AI提供商: auto, azure, openai, workers; auto为自动选择一个有效的配置，判断优先级为 azure > openai > workers                                           |
+| UPDATE_BRANCH             | 分支                                         | `master`                                                            | 版本检测所在分支                                                                                                                        |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| TELEGRAM_API_DOMAIN       | Telegram                                   | `https://api.telegram.org`                                          | 可以自定义Telegram服务器                                                                                                                |
+| TELEGRAM_AVAILABLE_TOKENS | 支持多个Telegram Bot Token                     | `null`                                                              | 多个Token用`,`分隔                                                                                                                   |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| CHAT_WHITE_LIST           | 聊天ID白名单                                    | `null`                                                              | 多个ID用`,`分隔，不知道ID，和机器人聊一句就能返回                                                                                                    |
+| I_AM_A_GENEROUS_PERSON    | 关闭白名单，允许所有人访问                              | `false`                                                             | 鉴于很多人不想设置白名单，或者不知道怎么获取ID，所以设置这个选项就能允许所有人访问， 值为`true`时生效                                                                         |
+| LOCK_USER_CONFIG_KEYS     | 锁定自定义用户配置                                  | `[]`                                                                | 可以锁定某些字段。比如设置为`CHAT_MODEL`就可以防止其他用户通过`/setenv`指令切换模型，多个字段用`,`分隔                                                                 |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| AUTO_TRIM_HISTORY         | 自动清理历史记录                                   | `true`                                                              | 为了避免4096字符限制，将消息删减                                                                                                              |
+| MAX_HISTORY_LENGTH        | 最大历史记录长度                                   | `20`                                                                | `AUTO_TRIM_HISTORY开启后` 为了避免4096字符限制，将消息删减                                                                                       |
+| MAX_TOKEN_LENGTH          | 最大历史token数量                                | 2048                                                                | 过长容易超时建议设定在一个合适的数字                                                                                                              |
+| GPT3_TOKENS_COUNT         | GTP计数模式                                    | `false`                                                             | 使用更加精准的token计数模式替代单纯判断字符串长度，但是容易超时                                                                                              |
+| GPT3_TOKENS_COUNT_REPO    | GPT3计数器资源所在Repo                            | `https://raw.githubusercontent.com/tbxark-arc/GPT-3-Encoder/master` | 加载 GPT3 Token 计数配置的资源文件                                                                                                         |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| SYSTEM_INIT_MESSAGE       | 系统初始化信息                                    | `你是一个得力的助手`                                                         | 默认机器人设定                                                                                                                         |
+| SYSTEM_INIT_MESSAGE_ROLE  | 系统初始化信息角色                                  | `system`                                                            | 默认机器人设定                                                                                                                         |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| ENABLE_USAGE_STATISTICS   | 开启使用统计                                     | `false`                                                             | 开启后，每次调用API都会记录到KV，可以通过`/usage`查看                                                                                               |
+| HIDE_COMMAND_BUTTONS      | 隐藏指令按钮                                     | `null`                                                              | 把想要隐藏的按钮写入用逗号分开`/start,/system`, 记得带上斜杠,修改之后得重新`init`                                                                           |
+| SHOW_REPLY_BUTTON         | 显示快捷回复按钮                                   | `false`                                                             | 显示快捷回复按钮                                                                                                                        |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| DEBUG_MODE                | 调试模式                                       | `false`                                                             | 目前可以把最新一条消息保存到KV方便调试,非常消耗KV写入量，正式环境务必关闭                                                                                         |
+| DEV_MODE                  | 开发模式                                       | `false`                                                             | 开发测试用                                                                                                                           |
+| STREAM_MODE               | 流模式                                        | `true`                                                              | 得到类似ChatGPT Web一样的打字机输出模式                                                                                                       |
+| SAFE_MODE                 | 安全模式                                       | `true`                                                              | 安全模式，会增加KV写损耗，但是能避免Workers超时导致的Telegram死亡循环重试，减少Token的浪费，不建议关闭。                                                                 |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| API_KEY                   | OpenAI API Key                             | `null`                                                              | 可以同时使用多个key，使用的时候会随机选择一个                                                                                                        |
+| CHAT_MODEL                | open ai 模型选择                               | `gpt-3.5-turbo`                                                     |                                                                                                                                 |
+| OPENAI_API_DOMAIN         | OPENAI API Domain [废弃: 使用 OPENAI_API_BASE] | `https://api.openai.com`                                            | 可以替换为其他与OpenAI API兼容的其他服务商的域名                                                                                                   |
+| OPENAI_API_DOMAIN         | OPENAI API Base URL                        | `https://api.openai.com/v1`                                         | 兼容Cloudflare AI 网关                                                                                                              |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| AZURE_API_KEY             | azure api key                              | `null`                                                              | 支持azure的API，两个密钥随便选一个就可以。如果你要默认使用azure，你可以设置`AI_PROVIDER`为`azure`                                                               |
+| AZURE_COMPLETIONS_API     | azure api url                              | `null`                                                              | 格式`https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/chat/completions?api-version=2023-05-15` |
+| AZURE_DALLE_API           | azure dalle api url                        | `null`                                                              | 格式`https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/dall-e-3/images/generations?api-version=2023-12-01-preview`   |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| CLOUDFLARE_ACCOUNT_ID     | Cloudflare 的 用户ID                          | `null`                                                              | 你可以在workers首页的右侧信息栏中找到这个信息。如果你要默认使用workers ai，你可以设置`AI_PROVIDER`为`workers`                                                      |
+| CLOUDFLARE_TOKEN          | Cloudflare的Token                           | `null`                                                              | 你可以在`https://dash.cloudflare.com/profile/api-tokens`中使用`Workers AI (Beta)`模板创建                                                  |
+| WORKERS_CHAT_MODEL        | 文字生成模型                                     | `@cf/meta/llama-2-7b-chat-fp16`                                     | 具体模型列表可以查看`https://developers.cloudflare.com/workers-ai/models/llm/`                                                            |
+| WORKERS_IMAGE_MODEL       | 文字生成图片模型                                   | `@cf/stabilityai/stable-diffusion-xl-base-1.0`                      | 同上                                                                                                                              |
+| -                         | -                                          | -                                                                   | -                                                                                                                               |
+| DALL_E_MODEL              | 生成图像的模型                                    | `dall-e-2`                                                          | 支持 `dall-e-2` 和 `dall-e-3`                                                                                                      |
+| DALL_E_IMAGE_SIZE         | 生成图像的尺寸                                    | `512x512`                                                           | 生成图像的尺寸。对于 dall-e-2，必须是256x256、512x512或1024x1024之一。对于 dall-e-3 模型，必须是1024x1024、1792x1024或1024x1792之一。                           |
+| DALL_E_IMAGE_QUALITY      | 生成图像的质量                                    | `standard`                                                          | 将要生成的图片质量。hd会创建具有更精细细节和整体一致性的图片。此参数仅支持dall-e-3.                                                                                 |
+| DALL_E_IMAGE_STYLE        | 生成图像的风格                                    | `vivid`                                                             | 生成图像的风格。必须是 vivid 或 natural 中的一个。vivid使模型倾向于产生超现实和戏剧化的图片。natural使模型产生更自然、不那么超现实外观的图片。此参数仅支持dall-e-3.                            |
 
 
 
@@ -88,22 +89,25 @@
 
 每个用户的自定义配置，只能通过Telegram发送消息来修改，消息格式为`/setenv KEY=VALUE`, 用户配置的优先级比系统配置的更高。如果想删除配置，请使用`/delenv KEY`。 批量设置变量请使用`/setenvs {"KEY1": "VALUE1", "KEY2": "VALUE2"}`
 
-| KEY                     | 说明                                                                                                               |
-|:------------------------|:-----------------------------------------------------------------------------------------------------------------|
-| AI_PROVIDER             | 配置与ENV相同                                                                                                         |
-| CHAT_MODEL              | 配置与ENV相同                                                                                                         |
-| OPENAI_API_KEY          | 设置该值之后将不会调用系统配置的KEY                                                                                              |
-| OPENAI_API_EXTRA_PARAMS | OpenAI API额外参数，设定后每次调用API都会带上，可以用来调整温度等参数， `/setenv OPENAI_API_EXTRA_PARAMS={"temperature": 0.5}`  每次修改必须为完整JSON |
-| SYSTEM_INIT_MESSAGE     | 配置与ENV相同                                                                                                         |
-| DALL_E_MODEL            | 配置与ENV相同                                                                                                         |
-| DALL_E_IMAGE_SIZE       | 配置与ENV相同                                                                                                         |
-| DALL_E_IMAGE_QUALITY    | 配置与ENV相同                                                                                                         |
-| DALL_E_IMAGE_STYLE      | 配置与ENV相同                                                                                                         |
-| AZURE_API_KEY           | 配置与ENV相同                                                                                                         |
-| AZURE_COMPLETIONS_API   | 配置与ENV相同                                                                                                         |
-| WORKERS_CHAT_MODEL      | 配置与ENV相同                                                                                                         |
-| WORKER_IMAGE_MODEL      | 配置与ENV相同                                                                                                         |
-
+| KEY                      | 说明                                                                                                               |
+|:-------------------------|:-----------------------------------------------------------------------------------------------------------------|
+| AI_PROVIDER              | 配置与  `ENV.AI_PROVIDER` 相同                                                                                        |
+| CHAT_MODEL               | 配置与  `ENV.CHAT_MODEL` 相同                                                                                         |
+| OPENAI_API_KEY           | 设置该值之后将不会调用系统配置的KEY                                                                                              |
+| OPENAI_API_EXTRA_PARAMS  | OpenAI API额外参数，设定后每次调用API都会带上，可以用来调整温度等参数， `/setenv OPENAI_API_EXTRA_PARAMS={"temperature": 0.5}`  每次修改必须为完整JSON |
+| SYSTEM_INIT_MESSAGE      | 配置与  `ENV.SYSTEM_INIT_MESSAGE` 相同                                                                                |
+| DALL_E_MODEL             | 配置与  `ENV.DALL_E_MODEL` 相同                                                                                       |
+| DALL_E_IMAGE_SIZE        | 配置与  `ENV.DALL_E_IMAGE_SIZE` 相同                                                                                  |
+| DALL_E_IMAGE_QUALITY     | 配置与  `ENV.DALL_E_IMAGE_QUALITY` 相同                                                                               |
+| DALL_E_IMAGE_STYLE       | 配置与  `ENV.DALL_E_IMAGE_STYLE` 相同                                                                                 |
+| AZURE_API_KEY            | 配置与  `ENV.AZURE_API_KEY` 相同                                                                                      |
+| AZURE_COMPLETIONS_API    | 配置与  `ENV.AZURE_COMPLETIONS_API` 相同                                                                              |
+| AZURE_DALLE_API          | 配置与  `ENV.AZURE_DALLE_API` 相同                                                                                    |
+| WORKERS_CHAT_MODEL       | 配置与  `ENV.WORKERS_CHAT_MODEL` 相同                                                                                 |
+| WORKER_IMAGE_MODEL       | 配置与  `ENV.WORKER_IMAGE_MODEL` 相同                                                                                 |
+| GOOGLE_API_KEY           | 配置与  `ENV.GOOGLE_API_KEY` 相同                                                                                     |
+| GOOGLE_COMPLETIONS_API   | 配置与  `ENV.GOOGLE_COMPLETIONS_API` 相同                                                                             |
+| GOOGLE_COMPLETIONS_MODEL | 配置与  `ENV.GOOGLE_COMPLETIONS_MODEL` 相同                                                                           |
 
 
 ### 支持命令
diff --git a/doc/en/CONFIG.md b/doc/en/CONFIG.md
index ac31d3b2..e8804736 100644
--- a/doc/en/CONFIG.md
+++ b/doc/en/CONFIG.md
@@ -15,54 +15,55 @@ It is recommended to fill in environment variables in the Workers configuration
 ### System Configuration
 Configuration that is common to each user, usually filled in the Workers configuration interface.
 
-| KEY                       | Description                                 | Default Value                                  | Special Description                                                                                                                                                                                                                                                                |
-|:--------------------------|---------------------------------------------|------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
-| AI_PROVIDER               | AI provider                                 | `auto`                                         | AI providers: auto, azure, openai, workers; auto automatically selects a valid configuration. The priority order is azure > openai > workers.                                                                                                                                      |
-| API_KEY                   | OpenAI API Key                              | `null`                                         | Multiple keys can be used at the same time, and one will be randomly selected when using the                                                                                                                                                                                       |
-| CHAT_MODEL                | Open AI model                               | `gpt-3.5-turbo`                                |                                                                                                                                                                                                                                                                                    |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| TELEGRAM_AVAILABLE_TOKENS | Support for multiple Telegram Bot Token     | `null`                                         | Multiple Token separated by `,`                                                                                                                                                                                                                                                    |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| CHAT_WHITE_LIST           | Chat ID Whitelisting                        | `null`                                         | Multiple IDs are separated by `,`, not knowing the IDs, talking to the bot for a sentence returns                                                                                                                                                                                  |
-| I_AM_A_GENEROUS_PERSON    | Close the whitelist and allow access to all | `false`                                        | Since many people don't want to whitelist, or don't know how to get an ID, setting this option will allow everyone to access it, with a value of `true`.                                                                                                                           |
-| LOCK_USER_CONFIG_KEYS     | Lock custom user configurations             | `[]`                                           | You can lock certain fields. For example, setting it to `CHAT_MODEL` can prevent other users from switching models through the `/setenv` command. Multiple fields are separated by `,`.                                                                                            |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| AUTO_TRIM_HISTORY         | Automatically trim history                  | `true`                                         | To avoid the 4096 character limit, truncate the message                                                                                                                                                                                                                            |
-| MAX_HISTORY_LENGTH        | Maximum history length                      | `20`                                           | `When AUTO_TRIM_HISTORY is turned on` To avoid the 4096 character limit, truncate the message                                                                                                                                                                                      |
-| MAX_TOKEN_LENGTH          | Maximum number of historical tokens         | 2048                                           | Too long and easy to timeout suggest setting at a suitable number                                                                                                                                                                                                                  |
-| GPT3_TOKENS_COUNT         | GTP counting mode                           | `false`                                        | Use more accurate token counting mode instead of just judging string length, but it's easy to time out                                                                                                                                                                             |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| SYSTEM_INIT_MESSAGE       | System initialization message               | `You are a useful assistant.`                  | Default robot init message                                                                                                                                                                                                                                                         |
-| SYSTEM_INIT_MESSAGE_ROLE  | System initialization message Role          | `system`                                       | Default robot init role                                                                                                                                                                                                                                                            |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| ENABLE_USAGE_STATISTICS   | Enable usage statistics                     | `false`                                        | After enabling, each API call will be recorded in KV and can be viewed through `/usage`.                                                                                                                                                                                           |
-| HIDE_COMMAND_BUTTONS      | Hide command buttons                        | `null`                                         | Write the buttons you want to hide separated by commas `/start,/system`, remember to include slashes, and after modifying, reinitialize `init`.                                                                                                                                    |
-| SHOW_REPLY_BUTTON         | Show Quick Reply button                     | `false`                                        | Display quick reply buttons.                                                                                                                                                                                                                                                       |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| UPDATE_BRANCH             | Git branch                                  | `master`                                       | Branch where version detection is located                                                                                                                                                                                                                                          |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| DEBUG_MODE                | Debug mode                                  | `false`                                        | Currently, the latest message can be saved to KV for convenient debugging. It consumes a lot of KV write volume and must be turned off in the production environment.                                                                                                              |
-| DEV_MODE                  | Developer mode                              | `false`                                        | Development testing                                                                                                                                                                                                                                                                |
-| STREAM_MODE               | Stream mode                                 | `true`                                         | Get a typewriter output mode similar to ChatGPT Web.                                                                                                                                                                                                                               |
-| SAFE_MODE                 | Safe mode                                   | `true`                                         | Safe mode will increase KV write overhead, but it can avoid Telegram's death loop retry caused by Workers timeout, reduce token waste, and it is not recommended to disable.                                                                                                       |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| LANGUAGE                  | Language                                    | `zh-CN`                                        | `zh-CN`，`zh-TW`, `en`                                                                                                                                                                                                                                                              |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| TELEGRAM_API_DOMAIN       | Telegram                                    | `https://api.telegram.org`                     | Customization of Telegram API server.                                                                                                                                                                                                                                              |
-| OPENAI_API_DOMAIN         | OpenAI                                      | `https://api.openai.com`                       | Can be replaced with the domain name of other service providers compatible with OpenAI API.                                                                                                                                                                                        |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| AZURE_API_KEY             | azure api key                               | `null`                                         | Support Azure API, choose either of the two keys. If you want to use Azure by default, you can set `AI_PROVIDER` to `azure`.                                                                                                                                                       |
-| AZURE_COMPLETIONS_API     | azure api url                               | `null`                                         | `https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/chat/completions?api-version=2023-05-15`                                                                                                                                                      |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| CLOUDFLARE_ACCOUNT_ID     | Cloudflare User ID                          | `null`                                         | You can find this information in the right sidebar of the workers homepage,  If you want to use Azure by default, you can set `AI_PROVIDER` to `workers`.                                                                                                                          |
-| CLOUDFLARE_TOKEN          | Cloudflare Token                            | `null`                                         | You can create using the `Workers AI (Beta)` template at `https://dash.cloudflare.com/profile/api-tokens`.                                                                                                                                                                         |
-| WORKERS_CHAT_MODEL        | Text generation model                       | `@cf/meta/llama-2-7b-chat-fp16`                | You can check the specific model list at `https://developers.cloudflare.com/workers-ai/models/llm/`.                                                                                                                                                                               |
-| WORKERS_IMAGE_MODEL       | Text-to-image generation model              | `@cf/stabilityai/stable-diffusion-xl-base-1.0` | Same as above.                                                                                                                                                                                                                                                                     |
-| -                         | -                                           | -                                              | -                                                                                                                                                                                                                                                                                  |
-| DALL_E_MODEL              | The model of the generated image            | `dall-e-2`                                     | Support `dall-e-2` and `dall-e-3`                                                                                                                                                                                                                                                  |
-| DALL_E_IMAGE_SIZE         | The size of the generated image             | `512x512`                                      | The size of the generated images. Must be one of 256x256, 512x512, or 1024x1024 for dall-e-2. Must be one of 1024x1024, 1792x1024, or 1024x1792 for dall-e-3 models.                                                                                                               |
-| DALL_E_IMAGE_QUALITY      | The image quality of the generated image    | `standard`                                     | The quality of the image that will be generated. hd creates images with finer details and greater consistency across the image. This param is only supported for dall-e-3.                                                                                                         |
-| DALL_E_IMAGE_STYLE        | The image style of the generated image      | `vivid`                                        | The style of the generated images. Must be one of vivid or natural. Vivid causes the model to lean towards generating hyper-real and dramatic images. Natural causes the model to produce more natural, less hyper-real looking images. This param is only supported for dall-e-3. |
+| KEY                       | Description                                           | Default Value                                                       | Special Description                                                                                                                                                                                                                                                                |
+|:--------------------------|-------------------------------------------------------|---------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| LANGUAGE                  | Language                                              | `zh-CN`                                                             | `zh-CN`，`zh-TW`, `en`                                                                                                                                                                                                                                                              |
+| AI_PROVIDER               | AI provider                                           | `auto`                                                              | AI providers: `auto, azure, openai, workers`; auto automatically selects a valid configuration. The priority order is azure > openai > workers.                                                                                                                                    |
+| UPDATE_BRANCH             | Git branch                                            | `master`                                                            | Branch where version detection is located                                                                                                                                                                                                                                          |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| TELEGRAM_API_DOMAIN       | Telegram                                              | `https://api.telegram.org`                                          | Customization of Telegram API server.                                                                                                                                                                                                                                              |
+| TELEGRAM_AVAILABLE_TOKENS | Support for multiple Telegram Bot Token               | `null`                                                              | Multiple Token separated by `,`                                                                                                                                                                                                                                                    |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| CHAT_WHITE_LIST           | Chat ID Whitelisting                                  | `null`                                                              | Multiple IDs are separated by `,`, not knowing the IDs, talking to the bot for a sentence returns                                                                                                                                                                                  |
+| I_AM_A_GENEROUS_PERSON    | Close the whitelist and allow access to all           | `false`                                                             | Since many people don't want to whitelist, or don't know how to get an ID, setting this option will allow everyone to access it, with a value of `true`.                                                                                                                           |
+| LOCK_USER_CONFIG_KEYS     | Lock custom user configurations                       | `[]`                                                                | You can lock certain fields. For example, setting it to `CHAT_MODEL` can prevent other users from switching models through the `/setenv` command. Multiple fields are separated by `,`.                                                                                            |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| AUTO_TRIM_HISTORY         | Automatically trim history                            | `true`                                                              | To avoid the 4096 character limit, truncate the message                                                                                                                                                                                                                            |
+| MAX_HISTORY_LENGTH        | Maximum history length                                | `20`                                                                | `When AUTO_TRIM_HISTORY is turned on` To avoid the 4096 character limit, truncate the message                                                                                                                                                                                      |
+| MAX_TOKEN_LENGTH          | Maximum number of historical tokens                   | 2048                                                                | Too long and easy to timeout suggest setting at a suitable number                                                                                                                                                                                                                  |
+| GPT3_TOKENS_COUNT         | GTP counting mode                                     | `false`                                                             | Use more accurate token counting mode instead of just judging string length, but it's easy to time out                                                                                                                                                                             |
+| GPT3_TOKENS_COUNT_REPO    | GPT3 Counter Resource Repo                            | `https://raw.githubusercontent.com/tbxark-arc/GPT-3-Encoder/master` | Resource file for loading GPT3 Token counting configurations                                                                                                                                                                                                                       |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| SYSTEM_INIT_MESSAGE       | System initialization message                         | `You are a useful assistant.`                                       | Default robot init message                                                                                                                                                                                                                                                         |
+| SYSTEM_INIT_MESSAGE_ROLE  | System initialization message Role                    | `system`                                                            | Default robot init role                                                                                                                                                                                                                                                            |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| ENABLE_USAGE_STATISTICS   | Enable usage statistics                               | `false`                                                             | After enabling, each API call will be recorded in KV and can be viewed through `/usage`.                                                                                                                                                                                           |
+| HIDE_COMMAND_BUTTONS      | Hide command buttons                                  | `null`                                                              | Write the buttons you want to hide separated by commas `/start,/system`, remember to include slashes, and after modifying, reinitialize `init`.                                                                                                                                    |
+| SHOW_REPLY_BUTTON         | Show Quick Reply button                               | `false`                                                             | Display quick reply buttons.                                                                                                                                                                                                                                                       |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| DEBUG_MODE                | Debug mode                                            | `false`                                                             | Currently, the latest message can be saved to KV for convenient debugging. It consumes a lot of KV write volume and must be turned off in the production environment.                                                                                                              |
+| DEV_MODE                  | Developer mode                                        | `false`                                                             | Development testing                                                                                                                                                                                                                                                                |
+| STREAM_MODE               | Stream mode                                           | `true`                                                              | Get a typewriter output mode similar to ChatGPT Web.                                                                                                                                                                                                                               |
+| SAFE_MODE                 | Safe mode                                             | `true`                                                              | Safe mode will increase KV write overhead, but it can avoid Telegram's death loop retry caused by Workers timeout, reduce token waste, and it is not recommended to disable.                                                                                                       |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| API_KEY                   | OpenAI API Key                                        | `null`                                                              | Multiple keys can be used at the same time, and one will be randomly selected when using the                                                                                                                                                                                       |
+| CHAT_MODEL                | Open AI model                                         | `gpt-3.5-turbo`                                                     |                                                                                                                                                                                                                                                                                    |
+| OPENAI_API_DOMAIN         | OpenAI  API Domain  [Deprecated: use OPENAI_API_BASE] | `https://api.openai.com`                                            | Can be replaced with the domain name of other service providers compatible with OpenAI API.                                                                                                                                                                                        |
+| OPENAI_API_BASE           | OPENAI API Base URL                                   | `https://api.openai.com/v1`                                         | Compatible with Cloudflare AI Gateway                                                                                                                                                                                                                                              |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| AZURE_API_KEY             | azure api key                                         | `null`                                                              | Support Azure API, choose either of the two keys. If you want to use Azure by default, you can set `AI_PROVIDER` to `azure`.                                                                                                                                                       |
+| AZURE_COMPLETIONS_API     | azure completions api url                             | `null`                                                              | `https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/chat/completions?api-version=2023-05-15`                                                                                                                                                      |
+| AZURE_DALLE_API           | azure dalle api url                                   | `null`                                                              | `https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/dall-e-3/images/generations?api-version=2023-12-01-preview`                                                                                                                                                        |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| CLOUDFLARE_ACCOUNT_ID     | Cloudflare User ID                                    | `null`                                                              | You can find this information in the right sidebar of the workers homepage,  If you want to use Azure by default, you can set `AI_PROVIDER` to `workers`.                                                                                                                          |
+| CLOUDFLARE_TOKEN          | Cloudflare Token                                      | `null`                                                              | You can create using the `Workers AI (Beta)` template at `https://dash.cloudflare.com/profile/api-tokens`.                                                                                                                                                                         |
+| WORKERS_CHAT_MODEL        | Text generation model                                 | `@cf/mistral/mistral-7b-instruct-v0.1 `                             | You can check the specific model list at `https://developers.cloudflare.com/workers-ai/models/llm/`.                                                                                                                                                                               |
+| WORKERS_IMAGE_MODEL       | Text-to-image generation model                        | `@cf/stabilityai/stable-diffusion-xl-base-1.0`                      | Same as above.                                                                                                                                                                                                                                                                     |
+| -                         | -                                                     | -                                                                   | -                                                                                                                                                                                                                                                                                  |
+| DALL_E_MODEL              | The model of the generated image                      | `dall-e-2`                                                          | Support `dall-e-2` and `dall-e-3`                                                                                                                                                                                                                                                  |
+| DALL_E_IMAGE_SIZE         | The size of the generated image                       | `512x512`                                                           | The size of the generated images. Must be one of 256x256, 512x512, or 1024x1024 for dall-e-2. Must be one of 1024x1024, 1792x1024, or 1024x1792 for dall-e-3 models.                                                                                                               |
+| DALL_E_IMAGE_QUALITY      | The image quality of the generated image              | `standard`                                                          | The quality of the image that will be generated. hd creates images with finer details and greater consistency across the image. This param is only supported for dall-e-3.                                                                                                         |
+| DALL_E_IMAGE_STYLE        | The image style of the generated image                | `vivid`                                                             | The style of the generated images. Must be one of vivid or natural. Vivid causes the model to lean towards generating hyper-real and dramatic images. Natural causes the model to produce more natural, less hyper-real looking images. This param is only supported for dall-e-3. |
 
 
 
@@ -89,23 +90,25 @@ You can add the bot to a group, and then everyone in the group can chat with the
 
 Each user's custom configuration can only be modified by sending a message through Telegram. The message format is `/setenv KEY=VALUE`. User configurations have higher priority than system configurations. If you want to delete a configuration, please use `/delenv KEY`. To set variables in bulk, please use `/setenvs {"KEY1": "VALUE1", "KEY2": "VALUE2"}`.
 
-| KEY                     | Description                                                                                                                                                                                                                                  |
-|:------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
-| AI_PROVIDER             | Configuration same as ENV                                                                                                                                                                                                                    |
-| CHAT_MODEL              | Configuration same as ENV                                                                                                                                                                                                                    |
-| OPENAI_API_KEY          | After setting this value, the system-configured KEY will not be called.                                                                                                                                                                      |
-| OPENAI_API_EXTRA_PARAMS | OpenAI API additional parameters, once set, will be included in every API call and can be used to adjust temperature and other parameters. `/setenv OPENAI_API_EXTRA_PARAMS={"temperature": 0.5}` Each modification must be a complete JSON. |
-| SYSTEM_INIT_MESSAGE     | Configuration same as ENV                                                                                                                                                                                                                    |
-| DALL_E_MODEL            | Configuration same as ENV                                                                                                                                                                                                                    |
-| DALL_E_IMAGE_SIZE       | Configuration same as ENV                                                                                                                                                                                                                    |
-| DALL_E_IMAGE_QUALITY    | Configuration same as ENV                                                                                                                                                                                                                    |
-| DALL_E_IMAGE_STYLE      | Configuration same as ENV                                                                                                                                                                                                                    |
-| AZURE_API_KEY           | Configuration same as ENV                                                                                                                                                                                                                    |
-| AZURE_COMPLETIONS_API   | Configuration same as ENV                                                                                                                                                                                                                    |
-| WORKERS_CHAT_MODEL      | Configuration same as ENV                                                                                                                                                                                                                    |
-| WORKER_IMAGE_MODEL      | Configuration same as ENV                                                                                                                                                                                                                    |
-
-
+| KEY                      | Description                                                                                                                                                                                                                                  |
+|:-------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| AI_PROVIDER              | Configuration same as  `ENV.AI_PROVIDER`                                                                                                                                                                                                     |
+| CHAT_MODEL               | Configuration same as  `ENV.CHAT_MODEL`                                                                                                                                                                                                      |
+| OPENAI_API_KEY           | After setting this value, the system-configured KEY will not be called.                                                                                                                                                                      |
+| OPENAI_API_EXTRA_PARAMS  | OpenAI API additional parameters, once set, will be included in every API call and can be used to adjust temperature and other parameters. `/setenv OPENAI_API_EXTRA_PARAMS={"temperature": 0.5}` Each modification must be a complete JSON. |
+| SYSTEM_INIT_MESSAGE      | Configuration same as  `ENV.SYSTEM_INIT_MESSAGE`                                                                                                                                                                                             |
+| DALL_E_MODEL             | Configuration same as  `ENV.DALL_E_MODEL`                                                                                                                                                                                                    |
+| DALL_E_IMAGE_SIZE        | Configuration same as  `ENV.DALL_E_IMAGE_SIZE`                                                                                                                                                                                               |
+| DALL_E_IMAGE_QUALITY     | Configuration same as  `ENV.DALL_E_IMAGE_QUALITY`                                                                                                                                                                                            |
+| DALL_E_IMAGE_STYLE       | Configuration same as  `ENV.DALL_E_IMAGE_STYLE`                                                                                                                                                                                              |
+| AZURE_API_KEY            | Configuration same as  `ENV.AZURE_API_KEY`                                                                                                                                                                                                   |
+| AZURE_COMPLETIONS_API    | Configuration same as  `ENV.AZURE_COMPLETIONS_API`                                                                                                                                                                                           |
+| AZURE_DALLE_API          | Configuration same as  `ENV.AZURE_DALLE_API`                                                                                                                                                                                                 |
+| WORKERS_CHAT_MODEL       | Configuration same as  `ENV.WORKERS_CHAT_MODEL`                                                                                                                                                                                              |
+| WORKER_IMAGE_MODEL       | Configuration same as  `ENV.WORKER_IMAGE_MODEL`                                                                                                                                                                                              |
+| GOOGLE_API_KEY           | Configuration same as  `ENV.GOOGLE_API_KEY`                                                                                                                                                                                                  |
+| GOOGLE_COMPLETIONS_API   | Configuration same as  `ENV.GOOGLE_COMPLETIONS_API`                                                                                                                                                                                          |
+| GOOGLE_COMPLETIONS_MODEL | Configuration same as  `ENV.GOOGLE_COMPLETIONS_MODEL`                                                                                                                                                                                        |
 
 
 ### Support command
diff --git a/src/env.js b/src/env.js
index a5a61c53..51c5d879 100644
--- a/src/env.js
+++ b/src/env.js
@@ -4,19 +4,35 @@ import './i18n/type.js';
  * @class Environment
  */
 class Environment {
+  // -- 版本数据 --
+  //
+  // 当前版本
+  BUILD_TIMESTAMP = process?.env?.BUILD_TIMESTAMP || 0;
+  // 当前版本 commit id
+  BUILD_VERSION = process?.env?.BUILD_VERSION || '';
+
+
+  // -- 基础配置 --
   /**
    * @type {I18n | null}
    */
   I18N = null;
+  // 多语言支持
   LANGUAGE = 'zh-cn';
-
-
+  // 检查更新的分支
+  UPDATE_BRANCH = 'master';
   // AI提供商: auto, openai, azure, workers, gemini
   AI_PROVIDER = 'auto';
 
+  // -- Telegram 相关 --
+  //
+  // Telegram API Domain
+  TELEGRAM_API_DOMAIN = 'https://api.telegram.org';
   // 允许访问的Telegram Token， 设置时以逗号分隔
   TELEGRAM_AVAILABLE_TOKENS = [];
 
+  // --  权限相关 --
+  //
   // 允许所有人使用
   I_AM_A_GENEROUS_PERSON = false;
   // 白名单
@@ -24,6 +40,8 @@ class Environment {
   // 用户配置
   LOCK_USER_CONFIG_KEYS = [];
 
+  // -- 群组相关 --
+  //
   // 允许访问的Telegram Token 对应的Bot Name， 设置时以逗号分隔
   TELEGRAM_BOT_NAME = [];
   // 群组白名单
@@ -33,10 +51,8 @@ class Environment {
   // 群组机器人共享模式,关闭后，一个群组只有一个会话和配置。开启的话群组的每个人都有自己的会话上下文
   GROUP_CHAT_BOT_SHARE_MODE = false;
 
-  // OpenAI API Key
-  API_KEY = [];
-  // OpenAI的模型名称
-  CHAT_MODEL = 'gpt-3.5-turbo';
+  // -- 历史记录相关 --
+  //
   // 为了避免4096字符限制，将消息删减
   AUTO_TRIM_HISTORY = true;
   // 最大历史记录长度
@@ -47,11 +63,27 @@ class Environment {
   GPT3_TOKENS_COUNT = false;
   // GPT3计数器资源地址
   GPT3_TOKENS_COUNT_REPO = 'https://raw.githubusercontent.com/tbxark-arc/GPT-3-Encoder/master';
+
+  // -- Prompt 相关 --
+  //
   // 全局默认初始化消息
   SYSTEM_INIT_MESSAGE = null;
   // 全局默认初始化消息角色
   SYSTEM_INIT_MESSAGE_ROLE = 'system';
 
+  // -- Open AI 配置 --
+  //
+  // OpenAI API Key
+  API_KEY = [];
+  // OpenAI的模型名称
+  CHAT_MODEL = 'gpt-3.5-turbo';
+  // OpenAI API Domain 可替换兼容openai api的其他服务商
+  OPENAI_API_DOMAIN = 'https://api.openai.com';
+  // OpenAI API BASE `https://api.openai.com/v1`
+  OPENAI_API_BASE = null;
+
+  // -- DALLE 配置 --
+  //
   // DALL-E的模型名称
   DALL_E_MODEL = 'dall-e-2';
   // DALL-E图片尺寸
@@ -61,6 +93,8 @@ class Environment {
   // DALL-E图片风格
   DALL_E_IMAGE_STYLE = 'vivid';
 
+  // -- 特性开关 --
+  //
   // 是否开启使用统计
   ENABLE_USAGE_STATISTICS = false;
   // 隐藏部分命令按钮
@@ -68,14 +102,8 @@ class Environment {
   // 显示快捷回复按钮
   SHOW_REPLY_BUTTON = false;
 
-
-  // 检查更新的分支
-  UPDATE_BRANCH = 'master';
-  // 当前版本
-  BUILD_TIMESTAMP = process?.env?.BUILD_TIMESTAMP || 0;
-  // 当前版本 commit id
-  BUILD_VERSION = process?.env?.BUILD_VERSION || '';
-
+  // -- 模式开关 --
+  //
   // 使用流模式
   STREAM_MODE = true;
   // 安全模式
@@ -85,13 +113,8 @@ class Environment {
   // 开发模式
   DEV_MODE = false;
 
-  // Telegram API Domain
-  TELEGRAM_API_DOMAIN = 'https://api.telegram.org';
-  // OpenAI API Domain 可替换兼容openai api的其他服务商
-  OPENAI_API_DOMAIN = 'https://api.openai.com';
-  // OpenAI API BASE `https://api.openai.com/v1`
-  OPENAI_API_BASE = null;
-
+  // -- AZURE 配置 --
+  //
   // Azure API Key
   AZURE_API_KEY = null;
   // Azure Completions API
@@ -104,7 +127,7 @@ class Environment {
   // Cloudflare Token
   CLOUDFLARE_TOKEN = null;
   // Text Generation Model
-  WORKERS_CHAT_MODEL = '@cf/meta/llama-2-7b-chat-fp16';
+  WORKERS_CHAT_MODEL = '@cf/mistral/mistral-7b-instruct-v0.1 ';
   // Text-to-Image Model
   WORKERS_IMAGE_MODEL = '@cf/stabilityai/stable-diffusion-xl-base-1.0';