← AgentToolbox · Browse tools

# QuoteProof — AgentToolbox

Need to check whether documentation says those exact words?
Check quotes against supported documentation, with source evidence.

Choose to inspect whether supported documentation contains a supplied quotation. Use match status and evidence positions; a match does not establish truth, and unknown-only output cannot qualify for payment.

Creator

AgentToolbox (`first_party`). Beta. Version `0.1.1`.

Cost

from $0.01 USDC per published success; USDC on Base, buyer chosen amount. Latency and agent token costs remain on failed outcomes.

Scope

Literal quotation fidelity · supported documentation hosts · not truth verification

Tool ID

quote-proof

Invocation

POST /v1/products/quote-proof/invoke

Success: At least one decisive quotation result with schema-validated evidence; unknown results remain explicit. Textual fidelity only, never claim truth.
Failure policy: Invalid input, failed criterion or unavailable decisive evidence never settles. Uncertain settlement requires reconciliation; never issue a replacement authorization.

## Inspect, test, pin, call

Full current contract JSON: https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof
Current success checks JSON and verifiable pin: https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof/criteria
Source snapshot success hash (verify current contract before authorizing): 1800eb994d37b99a0bf1e82398ff236ed006fcb1e44bf62b0501e018f70a3d95
Free static fixtures JSON: https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof/examples
Fixtures are synthetic offline cases, not live outcomes. input_valid covers the full runtime-accepted contract, including any host/subset/seed constraints.
No real-input preview is published for this tool.
Read-only minimum challenge/pin: GET https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof/invoke (402 expected)
Direct minimum call: POST https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof/invoke
Optional higher-amount quote: POST https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof/quote

Exact buyer procedure and receipt replay: https://agi.agenttoolbox2026.workers.dev/buy.md
Reviews: https://agi.agenttoolbox2026.workers.dev/products/quote-proof/reviews

## Example input

```json
{
  "urls": [
    "https://developers.cloudflare.com/workers/platform/limits/index.md"
  ],
  "quotes": [
    {
      "source_index": 0,
      "quote": "CPU time"
    }
  ]
}
```

## Runtime limits

```json
{
  "urls": 5,
  "quotes": 10,
  "source_bytes": 100000,
  "robots_bytes": 16384,
  "redirects": 2,
  "deadline_ms": 12000,
  "context_words_per_source": 200,
  "output_bytes": 14000,
  "supported_hosts": [
    "developers.cloudflare.com",
    "docs.payai.network",
    "docs.x402.org",
    "docs.python.org",
    "nodejs.org",
    "developer.mozilla.org",
    "docs.github.com",
    "www.typescriptlang.org"
  ],
  "preview": "Disabled: a verdict is the paid result."
}
```

No prepared-result preview is offered for this tool. The shared retention metadata below mentions preparations; for this tool, completed paid output replay lasts 24 hours and pending settlement retains output for reconciliation.

## Schemas and retention

Input and output schemas: https://agi.agenttoolbox2026.workers.dev/v1/products/quote-proof
{
"inputs": "Raw requests/signatures/capabilities are not stored. Hash commitments stay private.",
"results": "Prepared results private for 15 minutes unpaid; completed paid output replay for 24 hours. Pending settlement retains output for reconciliation.",
"source_content": "Untrusted data; never instructions."
}