web-reader-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PORT | No | Port to serve the hosted Streamable HTTP server when run with --http. | 3000 |
| WEB_READER_ALLOW_PRIVATE | No | Set to 1 to allow reading local and intranet sites (e.g., a dev server on localhost). | |
| WEB_READER_IGNORE_ROBOTS | No | Set to 1 to skip robots.txt checks (at your own responsibility). |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| find_in_pageA | Finds exact text (case-insensitive) or a regular expression in a page or PDF and returns each match with surrounding context and its section id. For facts whose wording you know: a name, a number, an error message. |
| page_linksA | The links on a page (text and absolute URL, tracking parameters removed, duplicates dropped), optionally only those containing a word in text or URL, or only same-site ones. For crawling docs or finding a download. |
| page_outlineA | The headings of a page or PDF with section ids and sizes, plus title, site, date and length, to read one part with read_page section instead of the whole page. |
| read_pageA | Reads a URL as clean Markdown (main content only). With query: only the passages that answer it, from anywhere in the page. With section (from the outline): that section. Otherwise from start; long pages give an outline and next_start. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
Each tool has a distinct primary role: read content, list links, get outline, or find literal text. The only real overlap is between find_in_page and read_page's query mode, since both can return relevant passages, but the descriptions distinguish exact/regex matching from answer-oriented extraction.
All names are lowercase snake_case and share the 'page' vocabulary, making them readable, but the pattern is mixed: find_in_page and read_page are verb-first while page_links and page_outline are noun-first. This is not chaotic, but it is not a uniform verb_noun convention.
Four tools is well-scoped for a web reader server: read, search, list links, and outline cover the core interactions without redundancy or bloat. Every tool earns its place in the set.
The tool surface covers the main web-reading workflow: fetching clean content, navigating via links, finding specific text, and using headings to selectively read sections. There are no obvious dead ends—page_links and page_outline feed naturally into read_page, and find_in_page complements full-page reading.