Skip to main content
Glama

AI structured data extraction from a web page

web_extract
Read-only

AI web scraping to structured JSON: the fields of a JSON Schema (object root, up to 4 KB) or a field-name list, extracted from one public web page or document (HTML, PDF, DOCX) fetched via SSRF-checked redirects. Returns data in the requested shape, null where the page states no value, a page quote per value when one matches the page text in code, final URL, title and truncation flag. AI-generated by Anthropic Claude under a fixed Kit Rx prompt; values can be wrong. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.

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": "web_extract"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlYeshttp or https URL of a public web page or document (PDF, DOCX, spreadsheet). Default ports only; credentials, IP addresses, localhost and private or single-label names are rejected.
fieldsNoComma-separated field names to extract as text, at most 25, each 1-40 of a-z, 0-9 and _ (e.g. height,location,opened). Give schema or fields.
schemaNoJSON Schema of the object to extract, URL-encoded JSON, at most 4 KB, object root. Supported: type, properties, items, enum, const, anyOf/oneOf, allOf of objects, local $ref with $defs, format, description. additionalProperties must be false or omitted; recursive schemas are rejected; length and range constraints are removed. At most 50 properties, 5 levels and 16 union-typed properties. Give schema or fields.
main_onlyNotrue (default) reads the page's main content only; false reads the whole page, including header, navigation and footer (where contact details often sit).

Schema Changelog

Changes observed during successful MCP inspections.

  1. Added

TDQS

A3.7/5.0
Behavior4/5

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

With only readOnlyHint and openWorldHint supplied, the description carries real weight and does: SSRF-checked redirects, no charge on failure or no-match, $0.005 USDC/x402 v2 pricing, output includes final URL, title and truncation flag, null for absent values, and a warning that AI-generated values can be wrong. The only blemish is the closing claim that a no-argument call returns a stored example answer, which conflicts with url being required and reads as boilerplate.

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?

The key capability leads and the dense single paragraph is largely load-bearing: output shape, pricing and model attribution all matter to an agent deciding to spend money. It loses a point for repeating 'AI-generated' twice and for the trailing no-argument sentence that muddies rather than clarifies.

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?

No output schema exists, so the description must describe returns, and it does: requested shape, null for absent values, page quote per value, final URL, title, truncation flag. It also covers cost, model, supported document types and failure billing, leaving only minor gaps such as timeout/size limits or what happens on partial extraction.

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 coverage is 100% and each parameter already documents limits, examples and the schema/fields alternation, so the baseline of 3 applies. The description mostly restates what the schema says (4 KB object-root schema, field-name list) and adds no new parameter-level meaning.

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

Purpose4/5

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

The description names a specific verb and resource: AI extraction of a JSON-Schema- or field-list-defined shape from one public web page or document (HTML, PDF, DOCX). That is concrete enough to distinguish it functionally from raw-content siblings like web_markdown or web_read, but no sibling is named and no explicit boundary is drawn, so it falls short of a 5.

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?

Usage is only implied: you call it when you want structured fields rather than page text, and it must be given either schema or fields. There is no explicit when-to-use/when-not guidance, no named alternative for plain-text or Q&A cases, and no guidance on choosing schema vs. fields beyond the schema's own 'Give schema or fields' note.

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.

Resources