Unofficial Midjourney-Web-API, for learning and research purposes only.
- ✨ Simple and easy to deploy
- 👋 Support for caching images, mainland China access friendly
- 💾 Support image Upscale function, get HD large image
- 📚 Multi-threaded concurrency, high-speed drawing
- 💻 Automatic database cleanup and error handling functions are perfect
- 🔐 Cross-domain restrictions can be set, anti-theft
- clone project
git clone https://github.com/CelestialRipple/Midjourney-Web-API
cd Midjourney-Web-API
2、 Get Cookie (make sure you can use Midjourney drawing in Discord)
- Go to the dialog box in Discord with Midjourney Bot -
- Open the developer tools of any browser (right click/F12) and check the network option
- Send any drawing request
- Search for interaction in Developer Tools and check the request header and load example: ! weixin ! weixin-2
- Fill the information in the request header into sender_params.json. It is worth noting that if you need multi-threaded drawing, please fill in the channelid array with the channelid of different channels (the maximum number of concurrency is 3 for Standard plan, 12 for PRO plan)
4、 Start Web-API
pip -r requirements.txt
python app.py
- (Optional) Go to app.py to configure cross-domain
nano app.py
- post request: http://localhost:5000/api/send_and_receive"
- Optional parameter: cdn=true (default false, after enabling the server will cache the image before sending, continental access is more friendly) Example:
import requests
import json
payload = {
"prompt": "your_prompt_here"
}
url = "http://localhost:5000/api/send_and_receive";
response = requests.post(url, data=json.dumps(payload), headers={'Content-Type': 'application/json'})
print(response.json())
- get request: http://localhost:5000/upscale"
- Mandatory parameter: file_name(string), the name of the file that needs to be executed upscale (e.g. rockgifinrock1971_link_and_zelda_33e8886f-adae-4579-9882-bae644005f5b.png)
- Mandatory parameter: number (number), the serial number of the image that needs to be executed upscale (example 1/2/3/4).
- Optional parameters: cdn=true (default false, after enabling the server will cache pictures before sending, continental access more friendly) Example:
import requests
base_url = 'http://localhost:5000' # Replace with the URL your Flask application is actually running on
file_name = 'rockgifinrock1971_link_and_zelda_33e8886f-adae-4579-9882-bae644005f5b.png' # Replace with your actual file name
number = 3 # Replace with the number you want to use
response = requests.get(f'{base_url}/upscale', params={'file_name': file_name, 'number': number})
if response.status_code == 200.
print('Success!')
print(response.json())
else.
print(f'Error: {response.status_code}')
print(response.text)
Q: How often is the information in sender_params.json updated? A: It has been two weeks since the project was run, and it is still not expired.
- Model switching
- Multi-account concurrency
- easier cookie retrieval
For suggestions and cooperation: [email protected] Ask the author to eat Crazy Thursday to speed up the project: https://afdian.net/a/hiripple/plan
MIT
Sender.py and Receiver.py are based on https://github.com/George-iam/Midjourney_api二次开发