Proofence — trust infrastructure for AI systems. See the platform

Developers

Proofence API

The Proofence API brings Proofence Engine verification into an application. Developers submit a claim or a proposed action and get back a trust score, grounded evidence and a policy evaluation. The API returns findings; deciding what to do with them stays with your system.

POST /v1/verify · illustrative
POST /v1/verify
{
  "input": "Model output or agent action",
  "context": { "sources": [...], "policies": ["disclosure"] }
}

// 200 OK  (values illustrative)
// Findings only: your system decides what to do next.
{
  "trust": { "level": "high", "explanation": "grounded in 3 sources" },
  "policy": { "evaluated": ["disclosure"], "concerns": [] },
  "evidence": [ { "source": "audit-log", "status": "verified" } ],
  "evidence_id": "ev_9a2c...",
  "signature": "..."
}

The Proofence API lets developers add verification to their own systems without rebuilding it. A request submits a claim or a proposed action; the response returns a trust score, the evidence behind it and an evaluation against the policies that apply. Because the same engine sits behind it, results from the API are consistent with the rest of Proofence.

The interface is designed to fit into existing pipelines. Verification can run where an AI output is generated, before it is stored, or ahead of an action, so checks happen where they matter rather than as a separate manual step. Responses are structured to be read by code, so your system can decide what to do with a finding.

The API informs; it does not authorize. Like the engine behind it, it never grants permission for an action. If you need an action stopped before it runs, that is the job of an enforcement point such as Proofence Gate, deployed in your environment.

Results are backed by the engine’s signed evidence, so a decision your system makes on the strength of an API response can be traced back to what the engine actually found. Access to the API is arranged directly with Proofence; get in touch to discuss it.

Capabilities

  • Verification endpoints

    Submit a claim and receive a trust score with grounded evidence, so verification becomes a call rather than a manual review.

  • Policy evaluation

    Evaluate a request against your organization’s rules and get back which rules applied and why.

  • Findings, not permission

    The response informs your system’s decision. Enforcement before execution belongs to Proofence Gate.

  • Pipeline-ready responses

    Returns structured results designed to slot into existing flows, so checks run at generation, storage or action time.

  • Signed evidence

    Results are backed by the engine’s signed evidence, so what your system acted on can be traced back later.

Bring verifiable trust to every AI decision.

Set up governance in AtlasProof by Proofence, or talk to us about verification, readiness assessment, or a Proofence Gate pilot.

Proofence API