generated: '2026-07-19'
method: searched
source: https://docs.inference.net/api/api-quickstart
docs: https://docs.inference.net
summary: >-
  Cross-cutting request/response semantics for the Inference.net API. The API is
  OpenAI-compatible: request and response envelopes mirror the OpenAI Chat
  Completions schema, so existing OpenAI SDKs and tooling work by pointing base_url
  at https://api.inference.net/v1.
authentication:
  style: bearer-api-key
  header: Authorization
  ref: authentication/inference-authentication.yml
compatibility:
  standard: openai-api
  base_url: https://api.inference.net/v1
  description: >-
    OpenAI-compatible endpoints (chat completions and related). Use the OpenAI SDK
    or plain HTTP.
capabilities:
  function_calling:
    supported: true
    docs: https://docs.inference.net/api/function-calling
  structured_outputs:
    supported: true
    docs: https://docs.inference.net/api/structured-outputs
  vision:
    supported: true
    docs: https://docs.inference.net/api/vision
async_inference:
  supported: true
  modes: [batch, group]
  docs: https://docs.inference.net/api/async-inference/overview
  description: >-
    Cost-effective asynchronous inference with flexible completion times; results
    delivered via webhooks (see asyncapi/inference-webhooks.yml).
rate_limit_signaling:
  scope: requests-per-minute
  ref: rate-limits/inference-rate-limits.yml
versioning:
  scheme: uri-path
  current: v1
  ref: lifecycle/inference-lifecycle.yml
data_retention:
  docs: https://docs.inference.net/api/data-retention
  default: Request data is not used for model training by default; encryption in
    transit and at rest; project-level retention controls.
notes:
- No documented idempotency-key contract was found; not asserted.