Free Ride is a cost-optimization utility that configures OpenClaw to use free AI models from OpenRouter. It automatically discovers, ranks, and configures the best available free models as your primary provider with intelligent fallback chains — ensuring you always have working AI access without incurring API costs.
Key Features
- Automatic model discovery and ranking that evaluates free models by quality metrics and selects the best available option as your primary
- Smart fallback configuration with ranked alternatives that activate when rate limits are hit, preventing service interruptions
- Configuration preservation that respects your existing OpenClaw settings while layering in free model routing
- Watcher daemon for continuous monitoring and automatic model rotation when better options become available
- OpenRouter integration using their intelligent free model picker as the first fallback for maximum reliability
Use Cases
- Running development and testing workflows without burning through paid API credits
- Maintaining always-available AI access for personal projects and experimentation
- Providing a cost-free baseline for teams evaluating different model providers
- Automating model selection so you always get the best free option without manual research
How It Works
The skill queries OpenRouter's API to identify all currently available free models, ranks them by quality metrics, and updates the OpenClaw configuration file (
~/.openclaw/openclaw.json
) with a primary model and ranked fallback chain. The first fallback is always OpenRouter's intelligent router, which automatically selects the best available free model when your primary hits rate limits.
Getting Started
Obtain a free API key from openrouter.ai/keys and set the environment variable
OPENROUTER_API_KEY
. Install the skill with
cd ~/.openclaw/workspace/skills/free-ride && pip install -e .
, then run
freeride auto
followed by
openclaw gateway restart
to activate your free model configuration.