Requstory MCP¶

Every business has a process nobody wrote down, and the person who knows it left the documentation to whoever answers the next question.

Server type:  Remote (Streamable HTTP), keyless public tools
Endpoint:     https://requstory.com/mcp
Auth:         none for public tools, API key for account tools
Tools:        10 at tools/list
Registry:     com.requstory/mcp
Docs:         requstory.com/developers/api
Repository:   github.com/3d7tech/requstory-cli (MIT, CLI)

Why This Matters for Operators¶

A company's own website is the only description of its business that it publishes in public. It says which teams exist, what the customer is supposed to do first, where a handoff happens and how many steps it takes before a stranger can buy something. Read carefully, it is an org chart and a process map written by accident.

This server reads it deliberately. map_website takes up to six pages of a public site and returns the customer journey, the teams with their owners, the processes and the handoffs between them. Then it does the part that matters: it names the teams with no owner, and the checks that competitors pass that this company fails. That is the difference between a research summary and a finding.

What It Returns, Verified¶

The endpoint is live and open: a bare POST initialize on https://requstory.com/mcp returned HTTP 200 with serverInfo name requstory, version 1.0.0, on protocol 2025-06-18, and a full tools/list of 10 tools. Three are public and need no key at all: search_companies, get_company_score and check_ai_readiness.

The published flow, from the server instructions:

search_companies finds companies Requstory has already mapped and scored, by name or website
get_company_score returns an ease-of-doing-business score out of 100 across four parts: find and understand, get started online, get help, and AI-agent readiness
check_ai_readiness tests whether agents can read, understand and act on any website: bots allowed, readable pages, product data, checkout, help
map_website maps the journey, teams, processes and handoffs from up to six pages, and the map is saved
get_map, list_maps and ask_map read a saved map and answer questions about it with citations, free from the map's pages and entries by default
get_ai_company builds a made-up company run end to end by agents from real public company maps, teams as agent lanes

Limits Worth Recording¶

Account tools (map_website job results, list_maps, get_map, ask_map, get_account) need an API key from the Requstory profile sent as a bearer token, and the free plan carries credits and a daily map allowance that get_account reports. ask_map answers from the map's own pages unless ai=true is set, which is the honest default and also the one to check before quoting an answer. The scores are computed from public pages only, so a company whose process lives behind a login is scored on what it shows a stranger, which is exactly what the score claims to measure.

FAQ¶

Does Requstory need an API key to be useful?¶

No. The three public tools, search_companies, get_company_score and check_ai_readiness, answer keyless, so an agent can score a prospect or audit a site's agent-readiness before anyone signs up.

What does the ease-of-doing-business score actually measure?¶

Four parts, all from public pages: whether a buyer can find and understand the company, whether they can start online, whether they can get help, and whether an AI agent can read and act on the site.

Can it map a company that is not in its index?¶

Yes. map_website works on any public website from up to six pages, and the result is saved to the account for get_map and ask_map.

Where does the data come from?¶

Public website pages only. The map is built from what the site publishes, and ask_map cites the map's own pages and entries rather than a model's memory.

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects