A clear path for Claude-compatible clients
Move a Claude integration with a documented Messages endpoint, explicit model discovery, and practical checks for streaming, tools, limits, and data handling.
Anthropic-compatible Messages API
Send Claude-style requests to the familiar /v1/messages endpoint. Keep your existing x-api-key, anthropic-version, model, and message fields while choosing an available model from the live catalog.
curl https://router.cheap/v1/messages \
-H "x-api-key: sk-your-api-key" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{"model":"claude-sonnet-5","max_tokens":64,"messages":[{"role":"user","content":"Reply with OK"}]}'What to verify before switching
- Confirm the exact model identifier and current token pricing in the live catalog.
- Test streaming, tool use, image inputs, and context limits with your own workload.
- Review retention, support, rate limits, and refund terms before sending sensitive data.
Keep the operational details explicit
Use environment variables for keys, set retry budgets, and preserve provider request IDs in your logs. Validate the model, context window, tool schema, and streaming behavior in staging before changing a production base URL.
See the security overview and the live service status before sending sensitive or time-critical workloads.
Frequently asked questions
Which endpoint does the Anthropic SDK call?
The compatible Messages endpoint is /v1/messages on router.cheap. Follow the exact headers and base URL in the current docs.
Are all Claude features guaranteed?
No provider gateway should imply that. Verify each model and capability in the live catalog and test your workload.
Can I compare Claude and GPT pricing here?
Yes. Use the live catalog and compare equivalent input/output token workloads rather than comparing model names alone.