Architect Inference
API by Liquidinference · rest
A live auction exchange for inference. Buying. Send an ordinary OpenAI- or Anthropic-shaped request. The exchange ranks the eligible standing offers by estimated charge, locks a firm price cap before the first token, and settles on measured use. The exchange block on the response says what was bought and what it cost. Send x-architect-explain: true to add a routing object that says how the exchange chose the offer. GET /v1/jobs/{job_id} returns the same block later. Purchasing rules go in the exchange object on the request body. They are a single optional object on the request root, rather than headers, so that an unmodified SDK can pass them through its extra_body option. The market. You can read the book, the router models and the depth of every market without a credential. These views do not name providers. The live feed names providers and requires a credential. Selling. A provider registers the deployments it serves using its own credential, and an…
Verification: observed. Source: crawl.
Capabilities (42)
- List the models in the Anthropic shape.
- The single-model view behind `/v1/anthropic/models/{id}` always in the…
- Buy music or a sound effect at auction.
- Buy spoken audio at auction in the OpenAI speech format.
- Buy a transcript at auction in the OpenAI transcription format.
- Buy one completion at auction in the OpenAI Chat Completions format.
- Buy embeddings at auction in the OpenAI embeddings format.
- List the collections and the models that each holds now.
- List the harnesses whose guide the exchange verified in the last 14 days.
- Return the jobs and tokens that the exchange has processed in total.
- List the models that have a list price.
- Report the depth of each market.
- List each served model that has a live offer with its price range and depth.
- List the resting offers without their providers.
- List the router models that a request can name in `model`.
- The live book over a WebSocket.
- Tokens and jobs bought through each target over a recent window. The targets…
- Edit images. This is the purchase of `/v1/images/generations` for clients…
- Buy images at auction in the OpenAI images format.
- Get what one job cost what it metered and how fast it ran.
- Buy one message at auction in the Anthropic Messages format.
- Estimate the input tokens of a Messages request.
- List the models that a request can name.
- Get one model router model or pinned `router@version`.
- List the settled jobs of this provider oldest first.