Prerequisites

• Python 3.12+
• On Windows, you need the Visual C++ Redistributable
• One of the following:
  • OpenAI API key
  • Mistral AI API key
  • Self-hosted LLM inference server with OpenAI-compatible API

Installation

FileChat is published on PyPI. Install it with your favorite package manager.

Here is an example for pip:

pip install filechat

Here is an example for uv:

uv tool install filechat

On Linux, you can make file indexing faster by running the local embedding model on your Intel Arc or NVIDIA GPU. Just install the xpu or cuda extra. For example:

pip install filechat[cuda]

Usage

Just run filechat like this:

filechat [project-directory]

When running for the first time, FileChat guides you through setting up your LLM provider.

Key Features

• Project Indexing: Automatically creates a searchable index of project files for easy access.
• Contextual Chat: Interact with an AI that comprehends your codebase, allowing for insightful discussions and queries related to your project.
• Real-time Updates: Detects and indexes changes in files instantaneously, ensuring the latest data is always accessible.
• Chat History: Retains a chat history akin to ChatGPT, organized by each directory for convenient reference.