Verify Truth Claim

Trigger verification of a pending truth claim using a specified method

Initiates the verification process for a truth claim. You can choose from multiple verification methods depending on the required assurance level and speed. The claim transitions from pending to verifying and a verification job is queued.

Verification results are delivered asynchronously. Poll the Get Truth Claim endpoint or subscribe to the truth_claim.verified webhook event to receive results.

Authentication

Requires X-API-Key header or Bearer JWT token. Tenant-scoped via X-Tenant-ID.

Path Parameters

id · string · required

The unique identifier of the truth claim to verify. Format: maip-tc:ULID.

Request Body

verification_method · string · required

The method to use for verification. One of: - witness — multi-agent witness consensus (fastest, fully automated) - ai — AI-powered analysis using document understanding and anomaly detection - human — routes to a human reviewer for manual assessment (highest assurance)

priority · string · default: "normal"

Processing priority for the verification job. One of: - normal — standard queue priority - high — expedited processing, prioritized in the verification queue

Response

id · string

The truth claim identifier.

status · string

Updated status. Transitions to verifying upon successful submission.

verification_id · string

Unique identifier for the verification job. Use this to correlate webhook events.

verification_method · string

The method being used for verification.

estimated_completion · string

ISO 8601 estimated completion timestamp. Accuracy varies by method: witness (minutes), ai (minutes to hours), human (hours to days).

Verification Methods

Method	Assurance	Speed	Description
witness	Medium	Minutes	Multi-agent quorum votes on claim validity
ai	Medium-High	Minutes-Hours	AI models analyze evidence and detect anomalies
human	Highest	Hours-Days	Human expert reviews claim and evidence

You can trigger multiple verification methods for the same claim to achieve layered assurance. Each method runs independently and results are aggregated.