Skip to content
Menu

MCP server / memory & knowledge / source-linked

Memory MCP Server

Persistent local knowledge graph with entities, observations, relations and JSONL storage. Use this record to separate documented interface facts from proposed evaluation steps, and preserve the checked date with the source links when the provider changes its setup.

Before you choose.

Identify the implementation and the job it supports.

Which implementation is this?
Persistent local knowledge graph with entities, observations, relations and JSONL storage. Use this record to separate documented interface facts from proposed evaluation steps, and preserve the checked date with the source links when the provider changes its setup.
Implementation source ↗

Persistent local knowledge graph with entities, observations, relations and JSONL storage. Use this record to separate documented interface facts from proposed evaluation steps, and preserve the checked date with the source links when the provider changes its setup.

Setup and connection.

The documented interface, authentication and installation path.

Connection / transport
stdio
Authentication
No API key; access to configured memory file.

Run the npm package and set MEMORY_FILE_PATH; establish naming and correction conventions before storing user information.

Setup reference ↗

Documented setup example. Review the source and credentials before running; this command was not executed for this article.

npx -y @modelcontextprotocol/server-memory

What it actually exposes.

Concrete capabilities with maintainer sources.

Documented capability / 01

Entities

Store typed nodes with observations.

Documentation ↗
Documented capability / 02

Relations

Create directed links between existing entities.

Documentation ↗
Documented capability / 03

Recall

Search nodes or open specific names.

Documentation ↗

Evidence and method.

What was checked and what was not tested.

WHO

Editorial record

Compiled by clawbot.ai.

HOW

Source method

Public maintainer documentation and repositories. This is not a hands-on product review.

WHEN

Last checked 2026-10-01

Setup examples are documented examples. Suggested trials and comparison advice are editorial guidance.

Sources support the adjacent factual statements. A source review does not establish uptime, performance, security or current account eligibility. Read the editorial policy.

Try a specific task.

A bounded evaluation you can reproduce in your own environment.

Maintain a project glossary (proposed; not tested)

Suggested evaluation, not an observed test result.

  1. Create fictional entities and one relation.
  2. Reconnect and ask for the relation.
  3. Correct one observation and inspect the graph.

What to inspect: Correction is retrievable and stale fact removed; proposed, not benchmarked.

Workflow reference ↗

Permissions and limitations.

Read these constraints before granting access.

Boundary / 02

Editorial inference: stored assertions have no automatic freshness or truth guarantee.

Boundary reference ↗

Decide by workflow.

Make the expected output explicit before you connect.

Correction is retrievable and stale fact removed; proposed, not benchmarked.

Keep the task inputs, tool calls, output and any errors together when you evaluate. The trial above is a proposal, not a recorded result.

Compare the alternatives.

Specific reasons to choose another route, with the tradeoff shown.

Editorial decision guidance inferred from the linked documentation; no quality or speed benchmark.
AlternativeChoose it whenTradeoff
Filesystem MCP ServerEditorial inference: you prefer readable durable notes.You define retrieval conventions.
Documentation ↗
Sequential Thinking MCP ServerEditorial inference: you need one-session plan revisions.It is not durable storage.
Documentation ↗

Questions before setup.

Answers specific to this implementation.

Where is data stored?

MEMORY_FILE_PATH selects the JSONL file.
Documentation ↗

Does it save full chats?

The documented model is facts and relations, not transcripts.
Documentation ↗