MachineReachBot

MachineReachBot is the identifiable diagnostic crawler used by MachineReach, a product of PRISMGRID TECHNOLOGIES PRIVATE LIMITED. It reads public web pages and public machine-readable metadata to prepare a user-requested visibility assessment.

What it fetches

• robots.txt and bounded sitemap entries;
• the requested page and a small set of relevant public pages;
• reviewed discovery paths for ARD draft catalogs, A2A Agent Cards, and OpenAPI descriptions;
• publicly linked protocol/package metadata.

It uses only passive GET or HEAD requests. It does not submit forms, authenticate, execute APIs, send A2A tasks, call MCP tools, execute JavaScript, or download and run code.

Request limits

• Up to 20 HTML pages and 40 total requests per scan by default.
• Maximum crawl depth 2; global concurrency 4; per-host concurrency 2.
• Per-resource timeout and 2 MB body limit.

How to block it

Add a specific rule to robots.txt. The scanner evaluates the specific bot group before wildcard rules.

User-agent: MachineReachBot
Disallow: /

Canonical production policy URL: https://machinereach.com/crawler. Outbound requests identify as MachineReachBot/<version> (+https://machinereach.com/crawler) when the production app URL is configured.

Blocked resources are reported as “Not evaluated — blocked by crawl policy,” not as missing or failed.

Contact and security

Support: [Add support email before launch]. Security reports: [Add security email before launch].