🌐 ChatFS Server

This is the server-side component for ChatFS.

It acts as a lightweight relay between:

• a web-based AI (Grok, Qwen, etc.)
• your local machine running the ChatFS client

---

👉 Looking for the actual tool?

You probably want the client repo:

➡️ https://github.com/deexor64/chatfs

That’s the part you run locally to share your workspace.

This repo is just the server that sits in between.

---

🧠 What this server does

• exposes simple URL-based endpoints for LLMs
• maintains WebSocket connections with clients
• forwards requests → client → response → back to LLM

It intentionally stays very thin. All filesystem logic and safety checks happen in the client.

---

🚀 Self-hosting (recommended)

You can run your own server instead of using a public one.

This is recommended if you care about:

• privacy
• control over sessions
• not sending requests through external infrastructure

---

⚙️ Requirements

• Python 3.12
• pip

---

📦 Setup (using PDM)

1. Clone the repo

git clone https://github.com/deexor64/chatfs-server.git
cd chatfs-server

---

2. Install dependencies

If you don’t have PDM:

pip install pdm

Then:

pdm install

---

3. Run the server

pdm run start

Server will start on:

http://0.0.0.0:8000

---

🐳 Deployment note (Render, etc.)

If your platform doesn’t have PDM preinstalled, use the provided build.sh:

bash build.sh

Then use this as your start command:

pdm run start

---

🔌 How it connects

• Client connects via WebSocket → /client/
• LLM sends requests → /{client_id}/...
• Server forwards everything to the correct client session

---

⚠️ Notes

• This server does not enforce filesystem safety
• All validation happens in the client
• Treat this as a relay layer, not a secure boundary

---

🔐 Privacy

If you use a public server, your requests pass through it.

For better privacy:

run your own instance and point the client to it