Evaluation record · mcp-server-memory

MCP Memory Server

v2026.7.4

Anthropic

MCPmemorystoragemcpmodel-context-protocol

Strong

About This MCP

MCP reference server providing persistent memory and knowledge graph capabilities: long-term retention, entity relationship tracking, and contextual recall across conversations. One of the seven reference servers still maintained after the 2025-05-29 archival; MCP governance moved to the Agentic AI Foundation (Linux Foundation) on 2025-12-09, latest stable spec 2025-11-25 (2026-07-28 revision at release-candidate stage). Critical for personalized AI but raises privacy concerns.

Last Evaluated: July 9, 2026

Official Website

Trust Vector Analysis

Dimension Breakdown

🚀Performance & Reliability

memory retrieval accuracy

Retrieval accuracy testing

Evidence

Vector Database Performance — Semantic search with vector embeddings provides good recall

mediumVerified: 2026-07-09

knowledge graph integrity

Graph consistency testing

Evidence

Graph Database Implementation — Entity relationships maintained with reasonable accuracy

mediumVerified: 2026-07-09

context recall

Recall quality assessment

Evidence

Memory System Testing — Can recall past conversations with varying accuracy depending on relevance

mediumVerified: 2026-07-09

storage scalability

Scalability testing

Evidence

Database Backend — Scales with underlying database (SQLite, PostgreSQL, etc.)

mediumVerified: 2026-07-09

query performance

Query latency testing

Evidence

Vector Search Performance — Fast retrieval for small to medium datasets (1-10ms typical)

mediumVerified: 2026-07-09

🛡️Security

access control

Access control testing

Evidence

Memory Isolation — User-level memory isolation depends on implementation

mediumVerified: 2026-07-09

data modification risk

Write operation risk assessment

Evidence

Memory Write Operations — AI can modify or delete stored memories

mediumVerified: 2026-07-09

memory poisoning risk

Data integrity risk assessment

Evidence

Security Analysis — AI can store incorrect or malicious information in memory

mediumVerified: 2026-07-09

query injection protection

Injection attack testing

Evidence

Database Query Security — Parameterized queries used but semantic search has different attack vectors

mediumVerified: 2026-07-09

audit logging

Logging capabilities assessment

Evidence

Memory Operations Logging — MCP protocol logs operations but detailed memory audit varies by implementation

mediumVerified: 2026-07-09

🔒Privacy & Compliance

pii storage risk

Privacy risk assessment

Evidence

Memory Content Analysis — Stores personal information, preferences, and conversation history indefinitely

highVerified: 2026-07-09

data retention control

Data retention controls review

Evidence

Memory Management — Manual memory deletion possible but no automatic retention policies

mediumVerified: 2026-07-09

consent management

Consent framework review

Evidence

Privacy Controls — No built-in consent mechanisms for memory storage

mediumVerified: 2026-07-09

right to deletion

GDPR right to erasure assessment

Evidence

Memory Deletion — Can delete specific memories but requires manual intervention

mediumVerified: 2026-07-09

embedding privacy

Embedding privacy analysis

Evidence

Vector Embeddings — Text converted to embeddings, sent to embedding provider (OpenAI, etc.)

mediumVerified: 2026-07-09

👁️Trust & Transparency

memory visibility

Memory transparency assessment

Evidence

Memory Inspection — Users can query and view stored memories

mediumVerified: 2026-07-09

knowledge graph explainability

Explainability assessment

Evidence

Graph Visualization — Entity relationships can be viewed but limited visualization tools

mediumVerified: 2026-07-09

documentation quality

Documentation completeness review

Evidence

Documentation — Community documentation with examples but could be more comprehensive

mediumVerified: 2026-07-09

open source transparency

Source code transparency review

Evidence

GitHub Repository — Open source implementation available for review

highVerified: 2026-07-09

⚙️Operational Excellence

ease of setup

Setup complexity assessment

Evidence

Setup Guide — Straightforward setup with database backend configuration

mediumVerified: 2026-07-09

storage efficiency

Storage efficiency testing

Evidence

Vector Storage — Efficient vector storage with compression options

mediumVerified: 2026-07-09

retrieval performance

Performance benchmarking

Evidence

Query Performance — Fast semantic search for typical workloads

mediumVerified: 2026-07-09

maintenance requirements

Maintenance overhead assessment

Evidence

Database Maintenance — Requires periodic cleanup and optimization

Anthropic - Donating MCP to the Agentic AI Foundation — memory is one of the seven reference servers still actively maintained after the 2025-05-29 archival; MCP governance moved to the Agentic AI Foundation under the Linux Foundation on 2025-12-09

npm - @modelcontextprotocol/server-memory — Latest release 2026.7.4 (published 2026-07-04); still listed among the seven maintained reference servers in modelcontextprotocol/servers README (repo ~88,300 stars, not archived)

mediumVerified: 2026-07-09

backup and recovery

Backup capabilities review

Evidence

Database Backup — Standard database backup procedures apply

mediumVerified: 2026-07-09

Strengths

+Enables true long-term memory and personalization across sessions
+Knowledge graph capabilities for entity relationship tracking
+Fast semantic search with vector embeddings
+Supports building contextual understanding over time
+Can dramatically improve AI assistant quality and relevance
+Open source implementation with flexibility

Limitations

!Significant privacy risk - stores personal information indefinitely
!No built-in PII detection or automatic data anonymization
!Embeddings typically sent to third-party providers (OpenAI, etc.)
!Limited consent management and data retention controls
!Memory poisoning risk - AI can store incorrect information
!GDPR/privacy compliance challenges without careful implementation
!STATUS 2026-07-09: actively maintained reference server (one of 7 retained after the 2025-05-29 archival); latest npm release 2026.7.4 (2026-07-04); no known CVEs against this server; governance under the Agentic AI Foundation (Linux Foundation, 2025-12-09)

Metadata

license: MIT

supported platforms

0: All platforms with database support

programming languages

0: TypeScript

1: Python

mcp version: 1.0

github repo: https://github.com/modelcontextprotocol/servers

github stars: 88259

storage backends

0: SQLite

1: PostgreSQL

2: Redis

3: Vector databases

embedding providers

0: OpenAI

1: Anthropic

2: Local models

vector dimensions: Varies (384-1536 typical)

first release: 2024-11

maintained by: MCP project (Agentic AI Foundation / Linux Foundation since 2025-12-09)

status: Official - Active

transport types

0: stdio

installation methods

0: npm

Use Case Ratings

code generation

Useful for remembering coding preferences and project context

customer support

Excellent for maintaining customer history and personalized support

content creation

Great for maintaining style preferences and project continuity

data analysis

Useful for remembering analysis patterns and user preferences

research assistant

Excellent for building knowledge graphs and tracking research progress

legal compliance

Privacy concerns with storing sensitive case information indefinitely

healthcare

High PHI storage risk; retention and consent management challenges

financial analysis

Risk of storing sensitive financial information long-term

education

Excellent for personalized learning and tracking student progress

creative writing

Outstanding for maintaining character details, plot threads, and story continuity