Skip to main content
Glama

ProofFetch — Cited Web Evidence

Map a site and get cited text (uses key credits)

proof_fetch_crawl

Maps a public site and returns cited text for its first pages (1 to 10 per call), one key credit per page that succeeds, with one fingerprint over the whole set. The rest of the map comes back as remainingUrls. Takes the same output options as proof_fetch_evidence. Needs the key as the MCP request header Authorization: Bearer . Example: {"url":"https://example.com/","limit":5,"formats":["markdown"]}.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "url": "<url>"
    },
    "name": "proof_fetch_crawl"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlYesAny page on the site, usually its home page, for example https://example.com/.
limitNoHow many pages to extract in this call.
imagesNoWith markdown: keep images as ![alt](src) (true, the default) or leave them out.
searchNoOptional: only addresses containing this text, for example "/docs/".
extractNoFill a JSON schema from the page's own structured data, no AI: each field is read from a CSS selector you give (selectors or the field's x-selector), else JSON-LD, meta tags or microdata. derived.extract.sources names where each value came from; fields with no source come back null and are listed in missing.
formatsNoOptional extra outputs in derived: markdown (clean Markdown with headings, lists, tables and links; the main content by default), links (every link on the page, absolute, marked internal or not), metadata (title, description, author, dates, Open Graph, JSON-LD). Each is fingerprinted (derived.sha256) and can be recomputed from the page bytes with the open verifier.
maxCharsNoMaximum normalized characters to extract. Free preflight returns metadata plus a preview of about the first 1,000 characters.
idempotencyKeyNoOptional. The same value on a retry within 7 days returns the stored pages without using credits again.
removeSelectorNoWith markdown: drop the parts of the page matching this CSS selector before reading, for example ".ads, .comments".
targetSelectorNoWith markdown: read only the parts of the page matching this CSS selector, for example "article" or "#pricing".
onlyMainContentNoWith markdown: leave out menus, headers, footers, cookie notices and sidebars (true, the default) or keep the whole page (false).

Schema Changelog

Changes observed during successful MCP inspections.

  1. Added

TDQS

A4.1/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Beyond the annotations it discloses the billing model (credit per successful page), the auth requirement, the all-or-nothing fingerprint over the returned set, and that overflow URLs come back as remainingUrls. The idempotencyKey behavior (same value within 7 days returns stored pages without new credits) adds nuance the idempotentHint=false annotation does not convey, though it sits in tension with that hint rather than contradicting it outright.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads what the tool does and what it costs in the first sentence, then auth and an example; four sentences with no filler. The phrase 'with one fingerprint over the whole set' is slightly compressed and could read ambiguously.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For an 11-parameter, nested-object tool with no output schema, the description supplies the return shape essentials (pages, per-set fingerprint, remainingUrls) plus cost and auth. It stops short of describing the derived/extract output structure, which is left entirely to the schema, but nothing needed to invoke it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the parameters are already well documented; the description only restates the limit range and points to proof_fetch_evidence for shared output options. The inline example payload adds a little concreteness but no semantics the schema lacks.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a concrete compound action (map a public site + return cited text for its first pages), the per-call scope (1 to 10 pages), and the salient constraint (public sites). This separates it from a map-only sibling like proof_fetch_map without the agent needing to open either schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Cost and auth conditions are given (one key credit per succeeding page, Authorization: Bearer header, optional idempotency key for retries), which is real usage guidance. However, it never says when to prefer this over proof_fetch_map, proof_fetch_evidence, or the free preflight alternative, so the agent must infer the routing itself.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.