CompanyGraph MCP server

• Registry nameio.companygraph/mental-modelverified in the MCP registry

The registry name sits under companygraph.io. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-10-26
• Tools14
• Server calls itselfcompanygraph-mcp-server
• Server version0.69.0
• Protocol version2025-06-18

On 2026-10-26 the server answered initialize and tools/list without a sign-in. First answer in 567 ms.

• Whether its tools do the product's job has not been judged yet.
• Checked on 2 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 2 so far
• Answeredmeasuring2 of the 7 daily checks it takes to print a share
• Median first answer574 ms
• Space in an agent's context≈ 2,894 tokensthe definitions of all its tools (14), which an agent carries on every turn
• Tool definitionsA · 93 of 100
• Typed results14 of 14 toolsdeclared outputSchema: results come typed, read once

• Parameters without a description: in 6 of 14

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• list_types

  Every type the model's schemas declare. Use first, to learn which types exist before listing or describing one. No input. Returns `types`, each with `type`, `name`, `tagline`, `owner` (the type it nests under, or null) and `count`, the…

• describe_schema

  One type's schema: its sections as written and its declared relations as data. Use to learn what an entity of the type may hold; for the whole vocabulary use describe_relations. Input: `type`. Returns `url`, the schema's own file…

• describe_relations

  What the schemas declare between types: references with form and cardinality, ownership, enums, joins, list kinds. Use for a diagram or audit; for one type's full schema use describe_schema. Optional `type` narrows every list, `direction`…

• list_rules

  The vocabulary's conventions the model is held to, R0 onwards, from its core's CONVENTIONS.md. Use to resolve a number a schema or refusal cites, such as R9. A company's own rules are entities of type `rule`: list_entities and get_entity…

• describe_rule

  One convention of the vocabulary as written, not a company's own rule (an entity of type `rule`: list_entities, get_entity). Input: `rule`, a number from list_rules such as R9, in either case. Returns `rule`, `title`, `part`, `text` and…

• list_checks

  The checks the model's own gate runs, from the checker release this server was built with. Use to see which rules a script enforces and which are left to a reader. No input. Returns `checks`, each with `name`, `rule`, `title` and what it…

• describe_errors

  What a refused call looks like. A tool's output schema covers its answers only, so use this once, to branch on refusals or validate them. No input. Returns `errors`, every `code` with `when` it is raised and the JSON Schema of its…

• list_entities

  Entities with their frontmatter's short facts, in address order or newest first. Use to browse, or for what runs on a date, came last or has a value; to find by words use search. Input: optional `type`, `owner`, `order`, `on`, `by`…

• get_entity

  One entity as structured data: fields, sections, tables, references both ways. Use to reason over one; for its page as written use fetch. Input: `id` or address, where its page sits, or `type` and `name`; an ambiguous name is refused with…

• list_references

  The model's edges, filtered and paged. Use to inspect one entity's relations or one kind of reference without taking whole entities. Optional `entity` (an id or address), `direction` (out, in, both; needs `entity`), `via`, `type` (the far…

• find_evidence

  Every edge into one skill, grouped by the type of the page that drew it. Use to check a claimed skill against its evidence. Input: `skill`, an id, canonical name or address; optional `limit`, `cursor`. Returns `skill`, `evidence` and…

• search

  Find entities by words, substring or exact name. `match: "words"` needs every query word's stem in name, tagline, fields, sections or cells; `"text"` (default) is a case-insensitive substring over the same; `"name"` the exact canonical…

• fetch

  One entity's page as written. Use to quote or display the source; for structured fields and references use get_entity. Input: `id` or address, from search, list_entities or any reference. Returns `id`, `title`, `type`, `url` and `text`…

• diagram

  A picture of the model as Mermaid; edge data: list_references. Input: `shape`; `id`, `domain`, `type` or `unit` narrow it. aggregate, flow, lifecycle take a context's `id`; organization without `id` draws the company; system a landscape…