wtf is chat_forge?

ChatForge is a self hosted chatbot platform built with Django on the backend and Groq for AI responses. It lets you set up separate projects, each with its own AI agent, and give each agent a custom system prompt that shapes how it responds. You can also upload files such as PDFs, Word documents, Excel sheets, PowerPoint slides, and various code or text files, and the agent will read their contents and use them as context when chatting with you. Getting started involves cloning the project, installing Python dependencies, adding a free Groq API key to a settings file, running database migrations, and starting the built in development server. From there you register an account, create a project, add a system prompt, optionally upload some files, and start chatting through a simple web interface built with plain HTML, CSS, and JavaScript. Under the hood, incoming chat requests pass through a layered backend: request validation, then loading the relevant project, prompts, files, and recent conversation history, followed by a service layer that talks to the Groq API. That service layer includes a rate limiter to stay within Groq's free tier limits, a circuit breaker that stops sending requests if Groq starts failing repeatedly, and automatic retries with backoff. Only a limited window of the most recent messages is sent to the AI on each request, which keeps costs and response times predictable even in long conversations. The project also includes a full authentication system using JWT tokens, ownership checks so users can only access their own projects, request rate limiting, CORS controls, and security headers, along with production settings like HTTPS enforcement when debug mode is off. This project suits developers who want a working example of a small production style chatbot platform, or anyone who wants to run their own private AI chat tool backed by uploaded documents rather than relying on a hosted service.