Aquanode LogoAquanode Docs
Endpoints

Calling an endpoint

Mint a token, call the endpoint's URL, and read the result back.

An endpoint is called the same way a job is triggered from a token: mint one on the endpoint, then send it in a header.

The URL and the token

Every endpoint has a URL, shown on its page. Mint a token there, then:

curl -X POST https://server.aquanode.io/api/v1/run/<endpoint-id> \
  -H "x-job-token: aq_job_..." \
  -H "content-type: application/json" \
  -d '{"inputs": {"prompt": "a cat"}}'

A token's entire vocabulary is "call this one endpoint and read the result back". It can reach nothing else on your account, which is what makes handing one to a caller outside your team safe.

Every token carries a rate limit and a run cap, both required at mint time. They defend against different things (hammering, and a slow bleed), and a token without both is how an account gets drained.

One call in, one response out

An endpoint's own app is called once per run: your server answers the request and that answer is the result, with nothing to poll on your app's side afterward. That is different from a workload like ComfyUI, which accepts a job and hands back an id you have to check on separately: an endpoint's image never needs that second step.

Read the call's own result the same way you read any run: GET /api/v1/run/<endpoint-id>/<run-id> with the same token.

A cold call still queues

If no instance is running, your call queues behind a rental the same as any other scale-up (see Scaling). The result isn't ready until that finishes.