OpenClaw
What is OpenClaw?
OpenClaw is an AI agent platform for conversational and coding workflows. When OpenClaw supports a generic OpenAI-compatible provider, you can point that provider at JoyToken and route model calls through JoyToken’s API key, policy, wallet, and model routing layer.
OpenClaw configuration keys can differ by version. Use the exact field names from your OpenClaw version; this page describes the provider contract JoyToken expects.
Setup
Step 1: Create a JoyToken API key
Create a dedicated key such as dev-openclaw.
Step 2: Set environment variables
Step 3: Add a JoyToken provider
If OpenClaw exposes a generic OpenAI-compatible provider, use this shape:
Step 4: Start OpenClaw
Restart OpenClaw after changing the provider config. New model calls should route through JoyToken.
Manual Configuration
If your OpenClaw configuration is agent-scoped, put the same provider contract inside the agent model config:
If OpenClaw supports request headers, add a request ID:
Model Format
Use auto during early evaluation. Pin models after quality and cost are stable for repeatable coding-agent workflows.
Monitoring Usage
Why use JoyToken with OpenClaw?
Multi-agent cost control
Use separate JoyToken keys for local testing, shared team agents, and unattended workflows.
Policy enforcement
Block disallowed models, tiers, IPs, or spend before an OpenClaw agent reaches an upstream provider.
Model routing
Keep OpenClaw configured with auto while JoyToken chooses the model inside policy.