AI & Memory MCP Server

Cognee

Name: Cognee
Author: topoteretes

Memory manager for AI apps and Agents using various graph and vector stores and allowing ingestion from 30+ data sources

Scan flagged something by topoteretes14,159Apache-2.0PythonIntermediatestdio

Safety Report

83% Scan flagged something

Scanned 5 months ago

5 passed 1 warnings

Security

Authentication

details

Optional API_TOKEN authentication for API mode; LLM_API_KEY required for operation

Pass

CORS Policy

details

API mode supports HTTP transport; CORS config not explicitly documented

Info

Rate Limiting

details

No built-in rate limiting documented; depends on underlying LLM API limits

Info

Known CVEs

details

No known CVEs; actively developed project

Info

Code Quality

Dependency Audit

details

Heavy ML/graph dependencies (LLM APIs, vector DBs, graph DBs); no published audit

Warning

Dangerous Patterns

details

No dangerous code patterns detected

Pass

License

details

Apache 2.0 License

Pass

Code Source

details

Open source on GitHub (topoteretes/cognee)

Pass

Community

Maintenance Status

details

Actively maintained; regular releases and community engagement

Pass

Compatibility

Claude CodeDocumented Claude integration via MCP stdio and API modes

Tested

CursorCursor integration documented in README

Tested

VS CodeStandard MCP config should work; not explicitly documented

Partial

WindsurfStandard MCP config should work; not explicitly documented

Partial

Claude DesktopClaude Desktop integration documented in official README

Tested

Tools Provided (11)

cognify

Turns your data into a structured knowledge graph and stores it in memory. Accepts text, documents, or other data inputs and builds queryable graph representations.

cognee_add_developer_rules

Ingests core developer rule files (e.g. coding standards, project conventions) into memory so they can be retrieved and enforced during development sessions.

codify

Analyses a code repository, builds a code graph capturing relationships between files, functions, and modules, and stores it in memory for intelligent code navigation.

delete

Removes specific data items from cognee datasets. Supports both soft deletion (mark as deleted) and hard deletion (permanently remove) modes.

get_developer_rules

Retrieves all developer rules that were previously ingested, generated from past interactions, and stored in memory.

list_data

Lists all datasets and their data items with their IDs, useful for inspecting stored memory contents before performing deletion or search operations.

save_interaction

Logs user-agent interactions and query-answer pairs to memory, enabling the system to learn from past conversations and improve future responses.

prune

Resets cognee completely, removing all stored data and graphs for a fresh start. Use with caution as this operation is irreversible.

search

Queries stored memory using multiple strategies: GRAPH_COMPLETION, RAG_COMPLETION, CODE, CHUNKS, SUMMARIES, CYPHER, and FEELING_LUCKY. Returns relevant information from the knowledge graph.

cognify_status

Tracks the progress of an ongoing cognify pipeline operation, reporting current processing stage and completion percentage.

codify_status

Tracks the progress of an ongoing codify (code graph building) pipeline operation.

Frequently Asked Questions

What is Cognee?

Memory manager for AI apps and Agents using various graph and vector stores and allowing ingestion from 30+ data sources

What tools does Cognee provide?

Cognee provides 11 tools including cognify, cognee_add_developer_rules, codify, delete. Each tool can be called by AI agents to perform specific operations.

Is Cognee safe to use?

We cannot tell you that, and nobody scanning at this scale can. What we did in March 2026 was shallow-clone Cognee and run five keyword searches over its source: auth identifiers, CORS configuration, rate-limit identifiers, the dependency manifest against a hardcoded list of 18 known-vulnerable packages, and unguarded eval/exec calls. At least one matched — the specific finding is in the scan report on this page. A text search cannot follow a variable, judge intent, or tell you what a server does once it is running, so "nothing matched" is not "safe". What the scan does and does not cover is written out at /mcp/security.

What are alternatives to Cognee?

Similar MCP servers include Context7, Hindsight, Basic Memory. Each serves a similar purpose but may differ in features, language, and compatibility.

Stay Updated on MCP Servers

New servers, safety alerts, and install guides — weekly.

Cognee

Install

Safety Report

Compatibility

Tools Provided (11)

Frequently Asked Questions

Similar MCP Servers

Context7

Hindsight

Basic Memory

Free AI writing tools

Stay Updated on MCP Servers