Your private ChatGPT

Q&A with business documents securely. Depend less on other teams for answers.

About The Project

Docq offers private and secure GenAI that unlocks knowledge from your organisation's documents, with minimal onboarding and operational effort.

Design tenets:

• Data stays within your boundary: bring ML models to the data. Cloud-vendor-hosted & self-hosted LLMs within the same cloud account to address data security and privacy concerns.
• Self-hosting: with your organisation's choice of cloud vendor. Minimal onboarding and operational effort. Can support air-gap use cases with local models.
• Multi-model: ability to utilise a variety of models and ability adopt new models quickly.
• Multi-modal: supports text today. Image, video, and audio formats are on the roadmap.
• Extensible: Abstractions to easily add custom data sources. Plugin system for extending application, from UI to database.

We offer Docq as a turnkey solution to lower the barrier for your organisation to adopt the latest AI innovation safely.

• Open-source (AGPLv3) to address your organisation's security and compliance needs.
• Commercial license and support available (via Docq.AI based in London, UK).
• Re-licensing option also available.

Vision

Be the goto OSS GenAI application stack for the privacy and security conscious.

Demo

A demo app is hosted by Streamlit Community Cloud, subject to frequent data refresh and feature update.

Credentials for admin user:

• username: docq
• password: Docq.AI

There are also recordings (Loom) to demonstrate Docq's major features:

• Ask your documents vs shared spaces
• Multi-user access

Deploy Your Own Instance

Deploy to Azure in 15mins

Click and follow the Azure wizard

For more details see section Deploy to Azure: 15 Minutes and Secure

Deploy to AWS in 15mins

  On the roadmap.

Deploy to GCP in 15mins

  On the roadmap.

Documentation

Documentation site is hosted by GitHub Pages.

• Overview
• User Guide
• Developer Guide

Roadmap

We aim to support the following features in the future:

• More data ingestion options such as SaaS data connectors and network storage options
• A plugin platform and ecosystem
• Model fine-tuning as an option for self-hosted, open-source LLMs

We also have plan to offer enterprise add-ons such as SSO support and enhanced audit.

Stay tuned!

Contact

Docq.AI - @docqai - [email protected]

Project Link: https://github.com/docqai/docq/

Licenses

The code in the repo is distributed under the licenses below. If a file has a specific license and copyright notice displayed then that wins.

• AGPLv3. See LICENSE and NOTICE for more information.
• Commercial licenses available via Docq.AI
• Copyright (C) 2023-2024 FABR Ltd t/a Docq.AI.

Made ❤️ London.