Call trust scoring directly from your pipeline or your agent.

A REST API and a native MCP server, both documented and ready today. Send a request from your pipeline, or give an agent the tool.

AccessREST API, unversioned pre-1.0, plus MCP over JSON-RPC
AuthBring your own key, documented at first integration

The /runs endpoint

The response includes every signal, its weight, and the compliance summary.

POST /api/runs { "brief": "...", "brand_id": "northwind", "file": <binary> } → { "run_id": "8841" } GET /api/runs/8841 { "iq_score": 0.78, "band": "partial", "signals": { "intent": { "score": 0.64, "weight": 0.30 }, "identity": { "score": 0.91, "weight": 0.25 }, "custody": { "score": 0.88, "weight": 0.20 }, "origin": { "score": 0.95, "weight": 0.15 }, "coherence": { "score": 0.82, "weight": 0.10 } }, "compliance": { "clauses_evaluated": 24, "passed": 23, "review": 1 } }

What the API gives you

A trust score your team can act on early

Wire the IQ Score into your own approval workflow: block, warn, or auto-pass based on a threshold you choose, so your reviewers see an issue at build time, with room to fix it, instead of after release. Call it again post-launch on the same asset to see what held up.

An agent can call it directly over MCP with no wrapper

MCP access means an agent can request a trust score directly, as a tool call, without a bespoke wrapper around the REST API.

One schema serves every consumer

The response shape you build your pipeline against is the exact shape the platform's own product surfaces read from, at any volume from a single run to a full enterprise pipeline.

Brand data reaches you through the same surface

The portable, versioned Brand profile behind Identity is reachable through the same API and MCP surface as a run, so your own tools read current brand truth directly, whether you're calling it once or wiring it into a pipeline for a team of 100,000.

Get API access

You get a key and the full schema reference to start building against.

Get early access