Basic
Usage basedA simple monthly balance for any model in the catalog.
$10per month
- $10 in routing credit every month
- Provider list pricing; credit applies first
- Access every currently available model
- Top up any time without changing plans
OpenAI-compatible routing infrastructure
Models keep changing.Your endpoint doesn’t.
Use GPT-5.6, DeepSeek, Kimi, Qwen, and more through one API. Provider failover is automatic, and usage logs exclude prompt bodies.
- OpenAI-compatible
- Usage-based billing
- Automatic provider failover
- No prompt bodies in usage logs
Keep the tools.
Change the route.
Bring tools that expose BYOK or OpenAI-compatible configuration. Use a dedicated setup guide where available; supported features vary by tool.
View the API referencePricing
One plan. Any live model.
Basic is $10/month and includes $10 in routing credit. Add more whenever you need it.
Trust by designRouting should be visible.
Routing should be visible.
Prompt logging shouldn’t be.
One endpoint with clear data handling, inspectable infrastructure, and automatic upstream recovery.
Auditable, AGPL-3.0 core
The routing backend is published under AGPL-3.0, so teams can inspect how requests, usage, and provider selection are handled.
No prompt bodies in usage logs
Request content is processed to reach the selected model, but prompt and completion bodies are not stored in usage logs.
Minimum operational metadata
Model, provider, token, latency, billing, and security signals are retained only where needed to operate and protect the service.
Provider-aware failover
Ordered provider chains and circuit breakers move eligible requests away from unhealthy upstream capacity automatically.
Ready when you are
One endpoint. Your next model.
Keep your OpenAI-compatible client. Swap the base URL, choose a model, and route your first request in minutes.