TRUSTBEFOREPAY · PROVIDER TRUST API · FREE · READ ONLY · X402 PRE-SPEND

# Provider Trust API for AI agents.

TrustBeforePay is SCRIPTMASTERLABS’ free read-only Provider Trust API for AI agents: it inspects identity, reachable runtime surfaces, payment-surface evidence, status signals and uncertainty before an agent trusts or spends, on x402 v2 with USDC on Base mainnet (eip155:8453), per the Provider Trust API page.

**Provider Trust inspects identity, reachable runtime surfaces, payment-surface evidence, status signals, and uncertainty before an agent trusts or spends.** It is read-only decision input, not an automatic allow-list, wallet signer, or permission to spend.

**TrustBeforePay** is SCRIPTMASTERLABS' read-only provider verification and reputation preflight for public HTTPS provider origins. It helps an AI agent inspect identity, reachable machine surfaces, payment metadata, status signals, and other provider evidence before the caller decides whether to connect, rely on, or spend with that service.

## How do you try the provider verification API?

Enter a public provider URL and select **Check provider** to run the free Provider Trust preflight.

Public provider URL Check provider

~~~
Ready.
~~~

## What is this useful for?

Use the Provider Trust API before adding an unfamiliar MCP server, x402 API, data provider, or agent endpoint to an automated workflow. It is especially useful when a caller needs to distinguish a real public service from stale directory metadata, compare declared versus reachable surfaces, inspect provider reputation evidence, or document why a provider was accepted into a tool chain.

## Why does this matter before x402 spend?

x402 can make an API machine-payable, but a payment challenge by itself does not prove the provider is the right provider for the job. A cautious agent can inspect the provider first, then separately inspect the selected resource’s fresh `PAYMENT-REQUIRED` terms, and only after that decide whether to authorize USDC.

For canonical SCRIPTMASTERLABS paid resources, current payment authority is x402 v2 with **USDC on Base mainnet** (`eip155:8453`). Provider Trust does not replace the fresh target challenge and does not turn provider evidence into permission to spend.

## How should you use the evidence?

Treat the result as decision input, not an automatic allow-list. A healthy endpoint can still return bad data later, and an unreachable endpoint may simply be temporarily unavailable. The caller remains responsible for policy, spending limits, permissions, and whether a provider is appropriate for the task.

Unknown evidence stays unknown. Missing evidence should not be silently converted into a positive or negative claim. This is intentionally different from a binary “safe/unsafe” badge that hides uncertainty.

## What are the machine endpoints?

Free Provider Trust API: `https://squeezeos-api.onrender.com/api/trust/provider?target= `

Primary Provider Trust MCP: `https://provider-trust-mcp.onrender.com/mcp`

The MCP surface provides a richer agent-native interface; the free API gives a quick origin-level provider verification preflight without requiring a payment.

## What is the useful boundary?

Provider Trust evaluates provider/source evidence. It does not claim that every statement produced by that provider is factually true, does not make irreversible trust decisions, does not sign a wallet, and does not authorize spending. For market-data fact verification use [Truth Verify](https://scriptmasterlabs.com/fact-verification-api).

[x402 Preflight →](https://scriptmasterlabs.com/x402-gateway) · [Fact Verification →](https://scriptmasterlabs.com/fact-verification-api) · [All API evidence →](https://scriptmasterlabs.com/agent-api-evidence)

## What did a live Provider Trust check measure on 2026-10-15?

**On 2026-10-15, ScriptMasterLabs' free Provider Trust check refused to vouch for example.com** — decision `insufficient_evidence`, reason code `TOO_FEW_VERIFIED_GATES`. The check ran live and free: `GET /api/trust/provider?target=https://example.com` returned HTTP 200 in 0.87 s with the full gate-by-gate evidence payload:

- **runtime_liveness: pass** — the origin answered (HTTP 200, text/html).

- **provider_identity: unknown** — no live Agent Card, ERC-8004 identity, or MCP server manifest observed (all three probes 404).

- **capability_observation, payment_rail_consistency, settlement_evidence_surface, historical_behavior_evidence: unknown** — no machine-readable evidence found on the origin.

That refusal is the product working as designed: one of six evidence gates passed, so the machine-readable answer is "not enough evidence" — never a guessed score. The check is free and read-only.

Method: one unauthenticated GET from an external client on 2026-10-15 against a neutral public target (example.com). The full JSON decision, gates, sources and observation timestamp are returned by the route itself.

## How does Provider Trust compare with a trust directory?

The result that currently owns the query "API provider trust verification service" in search is Clyro (ScriptMasterLabs authority watch, 2026-10-15). A directory lists and describes providers. TrustBeforePay does a different job: it runs a live evidence evaluation against the provider's own origin at call time and returns a machine-readable decision with per-gate evidence — including an explicit refusal, as the example.com run above shows, when the evidence is thin. Use a directory to discover providers; use Provider Trust before an agent relies on or pays one.

## Claim receipts

Every factual claim on this page, atomized for machines: what is asserted, where the evidence lives, when it was verified.

| CLAIM | EVIDENCE | VERIFIED |
| --- | --- | --- |
| TRUSTBEFOREPAY · PROVIDER TRUST API · FREE · READ ONLY · X402 PRE-SPEND | Evidence | 2026-09-27 |
| Provider Trust inspects identity, reachable runtime surfaces, payment-surface evidence, status signals, and uncertainty before an agent trusts or spends. It is read-only decision input, not an automatic allow-list, walle | Evidence | 2026-09-27 |
| Live run 2026-10-15 (free): the Provider Trust check on example.com returned HTTP 200 in 0.87 s with decision insufficient_evidence (TOO_FEW_VERIFIED_GATES) — runtime_liveness passed; provider_identity, capability_observation, payment_rail_consistency, settlement_evidence_surface and historical_behavior_evidence were unknown. | Evidence | 2026-10-15 |

---
Original HTML: https://scriptmasterlabs.com/provider-trust-api