Decisions from your own code.

• A request names a model, carries a state and asks questions, each under an id you choose. A state is a record of your model’s fields, or, for Helm’s general model, any text or JSON.
• A question is typed: noul is yes or no, answered as the chance of yes; choice is one of several options, answered with a chance for each; score, on Helm’s general model, is one of several ordered levels.
• Every answer comes back under its question’s id, and is sealed before the response is sent.

• helm-default is Helm’s general decision model. It answers any state and any typed question with no setup, and it is not calibrated to your decisions.
• Your own models are set up in the app, from a file of your past decisions and what happened or from a question you write. Each answers its own question, gives its reasons, and is scored as you record what happened.

31 operations are documented here, all under /v1.

• The contract is GET /v1/openapi.yaml: all of /v1 in OpenAPI, read with no key.
• /v1 only adds fields; breaking changes ship as /v2.
• Only /v1 is supported. Other routes can change without notice.

The rest of /v1 is left out of these pages: the research (17), jobs (5), walkthroughs (8) and cases (6) routes. Developer, inside the app, lists every route, and the contract describes them all.

• QuickstartA key, a first call to Helm’s general model, then a model of your own: forecasts and outcomes.
• API referenceEvery route of the Helm Decision API, with its method, path, scope and what it does.
• MCP connectorRead-only tools over JSON-RPC 2.0, for agents and model clients.
• WebhooksEvent types, how a delivery is signed, and the delivery log.
• Use your own modelHave Helm ask your own server for a model’s answers, then seal, score and calibrate them.
• LimitsRate, concurrency and the daily allowance, and the answer each limit gives.