Services
One API for your applications and coding tools. Call it in the OpenAI-compatible format or the Anthropic Messages format with the same key, and see exactly what every request costs.
OpenAI-compatible endpoints and the Anthropic Messages API are served side by side. The same key, limits and metering apply to both.
Responses stream token by token over server-sent events in both formats, with exact token counts for every call.
Create and revoke keys, and give each one its own monthly cap, request-rate limit and expiry. Secrets are shown once and stored hashed.
Every account has a monthly spend limit and a five-hour window, so a runaway script cannot use up a month's budget in an afternoon.
Requests, tokens and cost by day, model or key, plus a recent-calls list with the status, latency and cost of each request.
Sign in to the portal with your JianyueLab Passport account. There is no separate password to manage.
Try any model you have access to in the browser, with streaming replies, a system prompt and sampling controls.
Quickstarts in curl, Python and TypeScript, with your base URL and a live model id already filled in.
Generate ready-to-use config files for popular coding agents. The files read your key from an environment variable and never contain it.
Sign in to the portal, create an API key, and point your SDK at https://llm.jianyuelab.net/v1. Clients built for the OpenAI API and for the Anthropic Messages API both work, with no bridge of your own in between.
Before a request runs, it is checked against your account's limits and the models it may use; afterwards it is metered in tokens and cost. When a limit is reached, the API returns a clear error code instead of charging more.
Our team is available to answer questions and help you choose the right solution.