Agent Trust, Verification & Machine-Payable Data Utilities API

Checks and payment terms on this page refer to the endpoint above. Other listed endpoints have not been checked individually.

Base URLhttps://x402.synthexforge.com

• POST/v1/governance/settle$0.02Agent Spend Governor — settle a reserved lease with the actual amount spent and release the remainder. IDEMPOTENT on lease_id: replaying the same lease returns the recorded outcome and never double-counts spend. Settling an expired lease is explicit (documented), never a silent success. Caller supplies {lease_id, actual_amount (USD)}.
• POST/v1/utils/code-review-brief$0.15Produce a concise advisory code review highlighting issues, security flags, and suggestions.
• POST/v1/veritrack/attest$0.25Attest a track-record or performance claim into a signed, verifiable, tamper-evident receipt.
• POST/v1/compliance/check$1Classify an AI use-case against EU AI Act Annex III risk classes and list applicable obligations.
• POST/v1/sec-extract$0.05Extract structured company and SEC filing data for a stock ticker.
• POST/v1/utils/doc-convert$0.1Convert a document between formats (PDF to text, text to JSON or Markdown).
• POST/v1/utils/pii-redact$0.25Redact personally identifiable information (names, emails, phones, cards, addresses) from a document or text.
• POST/v1/security/dependency-risk$0.2Assess and pin the dependency and supply-chain risk of a software dependency tree with signed reproducible evidence.
• GET/api/ebook/download$29Download the full ebook 'x402 — The Agent Economy' (Facilitator Press / Synthex Forge): 121 pages, 12 chapters, 19 code blocks on building machine-native services agents pay for over x402. Free metadata at GET /api/ebook, free sample at GET /api/ebook/sample.
• POST/v1/readiness/attest$0.25Attest, from caller-supplied readiness evidence per a supplied policy snapshot, that a system/component/service is containment-ready. Wraps the live veritrack-attest spine (Ed25519 + SHA-256). Deterministic pure-function engine (readiness_kb.py): carries caller fields verbatim, rejects missing required fields, no lookahead/leaks/fabrication, NO kill-switch execution authority (compliant boundary). Verify free at GET /v1/readiness/attest/verify.
• POST/v1/identity/control-proof$0.2Attest, via a NON-DETERMINATIVE signed CONTROL-PROOF, that an entity documents CONTROL of an identifier (a key/DID/wallet), wrapped around the live veritrack-attest spine (Ed25519 + SHA-256). Deterministic pure-function engine (identity_kb.py): carries caller fields verbatim, rejects missing required fields, no lookahead/leaks/fabrication, no KYC / no identity determination (compliant boundary). Verify free at GET /v1/identity/control-proof/verify.
• POST/v1/security/redteam/authorized-scan$0.25Pay-per-assessment authorized/adversarial-testing-as-a-service MCP behind a PROVABLE authorization gate (Mercury cycle-0009 LANE-1). REFUSES to run unless the caller provably owns the target (DNS/TXT or .well-known token hash) OR holds a signed scoped permit — the gate is validated by hash equality against the caller-supplied expected-token hash, NO live probe/DNS/HTTP fetch is made. Caller supplies the target descriptor + authorized test categories + authorization evidence; the engine deterministically gates, notarizes the authorization evidence, and emits ONE Ed25519-signed, SHA-256-reproducible authorization-gated red-team envelope ({gate_outcome, compliance_grade, planned_tests[], attestation_id}). Every refusal is itself signed + anchored so the service can PROVE it only ever ran on authorization. This is a compliance/attestation layer, NOT an attack capability. Re-check free at POST /v1/security/redteam/recheck.
• POST/v1/security/threatmodel$0.25Pay-per-analysis MCP/A2A tool-permission threat model. Caller supplies the tool manifest (MCP tools/list or A2A skills description); the engine deterministically enumerates scopes+permissions, flags high-risk capabilities (payment/exec/net/storage/identity), maps escalation paths and MCP/A2A-specific attack vectors, grades overall risk (low/medium/high/critical) via a fixed scoring function, and emits ONE Ed25519-signed, SHA-256-reproducible threat-model document keyed by attestation_id. No live probe is made and nothing is fabricated — the same manifest re-analysed against the same anchor yields byte-exact the same threat model. Verify free at GET /v1/security/threatmodel/verify/{attestation_id}.
• GET/v1/signed-research$0.5Generate a tamper-evident signed research brief on a requested topic (Ed25519 + SHA-256).
• POST/v1/verify-catalog/register$0.25Register any x402 service's signed output into the PUBLIC network-effect VERIFY catalog: re-checks it (canonical SHA-256), honestly assesses the signature, counts network-effect edges, and grows an honest trust index (Ed25519 + SHA-256). Verified re-checkable free.
• POST/v1/bureau/reputation$0.1x402 Verification Bureau — reputation score for an endpoint or issuer computed deterministically from caller-supplied SIGNED observation history ({observed_at, verdict|outcome, ...}). Unverifiable events are excluded and NAMED in coverage_gaps — never silently counted. Dated, Ed25519-signed report, verifiable FREE at POST /v1/bureau/verify.
• POST/v1/evidence/query$0.15Bind a set of sealed forensics-vault entries (or an actor filter) into ONE signed, SHA-256-reproducible query receipt (Ed25519 + SHA-256).
• POST/v1/agent-reputation$0.25Compute + sign a deterministic reputation composite ('Credit Bureau of the Agentic Economy') for an agent from its signed outcome events.
• POST/v1/governance/reserve$0.02Agent Spend Governor — ATOMIC reserved lease. Caller supplies {agent_id, amount (USD), intent, ttl_seconds?}. The reservation is a single serialised DB transaction (BEGIN IMMEDIATE) that re-reads the live reserved total and REFUSES if a per-intent/session/daily cap would be exceeded — two concurrent reserves can never double-spend. Returns a signed, expiring lease (Ed25519, verifiable free). The adoption entry check route is FREE at POST /v1/governance/check.
• POST/v1/data/c2pa/verify$0.25Pay-per-verify C2PA content-provenance attestation. Caller supplies a content hash (64-char lowercase hex, 0x prefix optional) + a C2PA-style manifest claim; the deterministic pure-function engine (c2pa_kb.py) verifies manifest well-formedness, binds the content_hash against the manifest's declared asset hash, extracts creator/generator/software_agent/signature/timestamp/assertion/ingredient counts, checks structural integrity (manifest_version, has_signature, has_timestamp, content_binding_present), and grades the provenance into ONE Ed25519-signed, SHA-256-reproducible attestation keyed by attestation_id with verdict in {AUTHENTIC, SUSPICIOUS, MALFORMED} + an issue list. Optional expected_creator / expected_generator strings are compared verbatim. No live network call is made and nothing is fabricated — the same inputs re-analysed against the same anchor yield byte-exact the same attestation. Read back free at GET /v1/data/c2pa/attestation/{attestation_id}.
• POST/v1/budget/preflight$0.25Pay-per-decision agent-safety gate an agent calls BEFORE any spend. Caller supplies {agent_id, action_amount_usd, policy {per_tx_cap, daily_cap, spent_today}}; the engine deterministically evaluates the proposed amount against the policy and returns an Ed25519-signed, SHA-256-reproducible allow/deny with remaining_budget_usd. Safe-deny on malformed input; the gate never settles or moves funds (caller-attested budget state). Verify sig free.
• POST/v1/receipts/mint$0.25Mint a TYPED, signed attestation receipt for any tx/action — a structured per-transaction receipt schema distinct from flat notarization, wrapped around the live veritrack-attest spine (Ed25519 + SHA-256). Deterministic pure-function engine (receipt_kb.py): carries caller fields verbatim, rejects missing required fields, no lookahead/leaks/fabrication. Verify free.
• POST/v1/invoice-ocr$0.02Extract structured fields (vendor, amount, date, line items) from an uploaded invoice or receipt.
• POST/v1/mcp/threatmodel/tool-permission$0.25Pay-per-analysis MCP that threat-models the caller's OWN MCP/A2A manifest for permission-sprawl takeover risk — the root of agent takeover. Caller supplies the manifest (MCP tools/list or normalized A2A skills) + (recommended) an ownership proof; the engine deterministically scores {overallRiskScore, riskClassHistogram, sprawlFindings[], verdict} into ONE Ed25519-signed, SHA-256-reproducible risk envelope. Auth-gated: CRITICAL verdicts require an ownership-proof'd manifest. Risk verdict only — no remediation authority. Verify free at POST /v1/mcp/threatmodel/tool-permission/recheck.
• POST/v1/agent-reputation/trusted-route$0.3Rank counterparties by reputation (premium routing): deterministic signed trusted-route ranking (best first) over the supplied per-route outcome histories (Ed25519 + SHA-256).

A dash means the listing gives no price or method. Copy buttons copy the full URL.

Showing 25 of 40 listed endpoints. Next endpoints