Skip to content

polats/free-the-claw

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

free-the-claw

A self-hosted AI chat setup using openclaw with free NVIDIA NIM models via claude-code-free.

Before You Begin

Get a free NVIDIA NIM API key from https://build.nvidia.com/settings/api-keys

Quick Start with Codespaces

Open in GitHub Codespaces

The fastest way to try free-the-claw — no local install needed.

  1. Click the Open in GitHub Codespaces badge above
  2. When prompted, add your NVIDIA_NIM_API_KEY as a Codespaces secret
  3. Wait for the containers to build and start
  4. The gateway port (18789) auto-opens in your browser — log in with ?token=changeme

Codespaces is ephemeral and will reset after a while. To persist your data, follow the Deploy on Railway section instead.

Running Locally Via Docker

  1. Clone the repo:

    git clone --recurse-submodules https://github.com/polats/free-the-claw.git
    cd free-the-claw
    
  2. Copy the example env file and add your API key:

    cp .env.example .env
    

    Edit .env and set your API key and model:

    NVIDIA_NIM_API_KEY=your-key-here
    MODEL=stepfun-ai/step-3.5-flash
    
  3. Start the services:

    docker compose up -d
    
  4. Open http://localhost:18789/?token=changeme in your browser.

Configuration

  • Model: Change MODEL in .env (default: stepfun-ai/step-3.5-flash)

Popular choices:

  • stepfun-ai/step-3.5-flash
  • moonshotai/kimi-k2.5
  • z-ai/glm4.7
  • minimaxai/minimax-m2.1
  • mistralai/devstral-2-123b-instruct-2512

See Claude Code Free to add more models.

  • Gateway token: Change OPENCLAW_GATEWAY_TOKEN in docker-compose.yml (default: changeme)

  • Ports: claude-code-free runs on 8082, openclaw on 18789

Deploy on Railway

Deploy on Railway

For persistent hosting, deploy on Railway with two services from this repo. You will probably need the Hobby Plan ($5/month) as the free tier may not have enough memory for openclaw.

  1. Click the Deploy on Railway badge above
  2. Add your NVIDIA NIM API key from https://build.nvidia.com/settings/api-keys in the openclaw service's variables
  3. Once deployed, you should see the gateway URL on the service (similar to https://openclaw-railway-1234567890.up.railway.app)
  4. Visit the URL via the browser appended with ?token=changeme (you can change this OPENCLAW_GATEWAY_TOKEN in the openclaw service's environment variables)

GitHub Integration (optional)

Connect your GitHub account by adding a Personal Access Token to your .env:

GITHUB_PAT_TOKEN=ghp_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

This does two things:

  1. Agent persona — auto-detects your GitHub username and fetches your profile README (the username/username repo) as a SOUL.md, which openclaw injects into the agent's system prompt
  2. Git operations — configures git credentials and gh CLI so the agent can clone, push, pull, and interact with your repos

About

Run Openclaw for free using NVIDIA NIM

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors