This project provides a starter application for building document querying systems powered by Large Language Models (LLMs). It targets developers and researchers looking to integrate LLM capabilities, such as text analysis and manipulation, into Django-based applications, offering a proof-of-concept for streaming LLM responses.

How It Works

Delphic leverages LlamaIndex and Langchain to build LLM agents for document analysis. The core approach involves using OpenAI's API for text processing, with plans to support other LLMs. This design choice prioritizes ease of deployment and access to capable models, though it currently mandates OpenAI API usage and its associated terms.

Quick Start & Requirements

• Install/Run: Clone the repo, create directories and copy sample environment files (mkdir -p ./.envs/.local/, cp -a ./docs/sample_envs/local/.frontend ./frontend, cp -a ./docs/sample_envs/local/.django ./.envs/.local, cp -a ./docs/sample_envs/local/.postgres ./.envs/.local), update .django with your OPENAI_API_KEY, then run sudo docker-compose --profile fullstack -f local.yml build and sudo docker-compose --profile fullstack -f local.yml up.
• Prerequisites: Docker, Docker Compose, Python >= 3.10 (recommended for development), Node v18.15.0 (recommended for frontend development).
• Setup Time: Estimated to be under 30 minutes for initial setup with Docker.
• Links: Docker official instructions

Highlighted Details

• Built with Django, LlamaIndex, and Langchain.
• Supports streaming responses from LLMs.
• Requires an OpenAI API key for current functionality.
• Docker-based deployment for ease of setup.

Maintenance & Community

The project is marked as not maintained, with a recommendation to check out OpenContracts for a more modern version. The author is seeking someone to take over maintenance.

Licensing & Compatibility

• License: MIT
• Compatibility: The MIT license permits commercial use and linking with closed-source projects. However, the current reliance on OpenAI's API means users must adhere to OpenAI's terms of service.

Limitations & Caveats

The application is not production-ready and requires significant upgrades. Currently, all logged-in users have full permissions, with plans to implement role-based access control. Collection creation and querying incur costs via OpenAI API credits.