MCP.Market

Memex MCP server

by evan-moon·io.github.evan-moon/memex·v0.5.0·3 stars
No reviews yet

Local-first second brain with semantic search. Markdown notes indexed by a local ML model.

C66/100grade C
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
C66/100

full report

Adoption
Growing

3 stars211 downloads/wk

Reviews

Write one

Nobody has reviewed Memex yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Memex tools (13, 3 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • delete_notewrite action

    Delete a note by ID. Removes both the markdown file and the index entry. Rule-layer notes are user-only and always rejected \u2014 suggest `memex delete <id>` instead.

  • get_inference

    Get one inference with full provenance: its summary, confidence, and every evidence note (with a snapshot of the note at mint time and whether it has since changed or been deleted). …

  • get_note

    Get the full content of a note by ID.

  • get_signals

    List deterministic "signals" \u2014 un-synthesized patterns the engine found in the notes (NO LLM involved, these are facts about the corpus structure): - hidden_arc: notes that clearly belong together but were never linked, spread over a long time (an un-synthesized thread) - stale_state: a "state" note that has newer related notes since it was last updated - dangling_link: a [[link]] to a note t

  • list_folders

    List all top-level folders (categories) in the second brain with their note counts. Call this before saving a note to pick the right folder.

  • list_inferences

    List inferences (LLM-synthesized hypotheses derived from notes). Staleness is re-checked first, so an inference whose source notes changed shows status "stale". …

  • list_notes

    List recent notes from the second brain.

  • list_tags

    List all tags in the second brain with their note counts. Call this before saving a note to pick consistent existing tags rather than inventing new ones.

  • mint_inference

    Persist a synthesized inference (a non-obvious claim that holds across several notes but is stated in none). ONLY call this when the user EXPLICITLY asks to save/record/keep a discovery ("save that", "record this insight"). NEVER auto-mint in the background \u2014 unvalidated hypotheses pollute the brain. First present the hypothesis in chat; mint only after the user approves. Provide either fromS

  • save_note

    Save a note to the second brain. Use proactively \u2014 without asking the user \u2014 at the end of any conversation that contains: technical decisions and their rationale, key points from meetings or 1-on-1s with specific people, newly learned concepts or insights, or project context worth recalling later. Prefer updating an existing note over creating a duplicate. `layer` is REQUIRED \u2014 cla

  • search_notes

    Search the second brain for relevant context. Call this BEFORE answering any question that could relate to past conversations, people, projects, or decisions the user may have stored. Always search first, then answer \u2014 even if the connection seems loose. For important or vague questions, search MORE THAN ONCE with different phrasings: once in the user's language and once in English, or once w

  • update_notewrite action

    Extend or correct an existing note. Use when new information belongs with an existing note rather than standing alone. Search first to find related notes, then update rather than creating a duplicate. Layer rules: - past notes are immutable \u2014 this tool will reject with PAST_IMMUTABLE. Create an [Amendment] save_note with a [[backlink]] instead. - rule notes are user-only \u2014 this tool will

  • update_signal_statuswrite action

    Triage a signal: dismiss noise so it never resurfaces, or snooze it for later. Use when a signal is clearly not worth synthesizing (e.g. a burst of boilerplate). Confirm with the user before dismissing.

Public scan report

scanner v0.1.9 · 2026-09-25 · same rubric, same numbers if you re-run it

1 high1 medium
  • Code scan5 source files scanned8/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitylocal package, no credentials required12/15
  • Maintenancelast push 5 days ago15/15
  • Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10

Findings (2)

  • highShell command built from a string (injection risk)exec.shell-concat
    dist/index.js: …X_FILES = 50; var git = (repo, args) => execSync(`git ${args}`, { cwd: repo, encoding: "utf8" })…
  • mediumeval / new Function usedexec.eval
    dist/mcp.js: …&&(d=this.opts.code.process(d,t));let m=new Function(`${ft.default.self}`,`${ft.default.scope…
Overall 66/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How Memex compares

Against every graded AI and agents MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
66/100
higher than 28% of 4,460 AI and agents servers
Tools
13
the typical AI and agents server has 6
Tools that change things
23%
typical: 0%. Fewer write tools means less an agent can break
GitHub stars
3
more than 75% of 2,711 AI and agents servers
Downloads a week
211
more than 75% of 1,538 AI and agents servers with a package

What its tools let an agent do: 9 read · 1 create · 2 change · 1 delete

Install Memex in Claude Code, Cursor or VS Code

Runs npx -y @evan-moon/memex on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add memex -- npx -y @evan-moon/memex
Add to Cursor

What the publisher says

From the Memex repository's README, as published. We do not edit it. Read it on GitHub

memex

Make Claude smarter about you.

npm install -g @evan-moon/memex

Make Claude smarter about you.

Local-first second brain that connects to Claude via MCP. Notes are stored as plain Markdown and indexed with a local ML model, fully offline, no API keys, nothing leaves your machine.

The problem

Claude is only as smart as what's in the conversation. Your decisions, your context, your thinking, invisible unless you paste them in every time.

You:    What did we decide about the auth approach last sprint?
Claude: I don't have context from previous conversations...

The fix

You:    What did we decide about the auth approach last sprint?

Claude: [memex · search_notes · "auth approach decision"]

        Found 2 notes:

        Auth Architecture Decision  Apr 14  #auth #backend
        ─────────────────────────────────────────────────────
        Chose JWT + refresh tokens over sessions. Rationale:
        stateless design fits horizontal scaling plan.

        Based on your April 14th note: you went with JWT +
        refresh tokens. Tom also flagged keeping auth decoupled
        from payment logic, separate bounded contexts.

Claude searches your notes before answering and saves insights at the end of every conversation, automatically, without being asked.

Install

npm install -g @evan-moon/memex

Connect your apps:

memex mcp install

This registers memex with every MCP client on this machine — Claude, Claude Code, Codex, Cursor — by writing each one's own config file. Restart the clients afterwards.

That's it. On first run, the embedding model (~450MB) downloads once to ~/.memex/models/.

Auto-recall (optional)

memex recall install

Turns retrieval from something Claude has to decide to do into something that just happens. Every prompt you type is semantically searched against your notes, and the top 3 titles are injected as context before Claude answers — the same way native memory works. Claude then pulls full notes with get_note when a title looks relevant.

A background daemon keeps the embedding model warm (~/.memex/recall.sock), so a lookup costs ~30ms instead of the ~1.5s a cold CLI search spends loading the model. It idles out after 2 hours.

Cost: ~200MB resident while warm, plus up to 3 note titles of context per prompt. Remove with memex recall uninstall.

Features

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Memex: common questions

Is Memex MCP server safe?
With care: it is graded C, so read the findings first (66/100). Read the Memex safety report
How do I install Memex?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Memex need an API key?
Not as far as the registry entry and our scan can tell: no credentials are declared or required.
Is Memex maintained?
The last commit was 45 days ago (2026-09-20). The latest release is v0.5.0.
What can I use instead of Memex?
Servers from other publishers that do the same job: Basic Memory MCP server, IWE (iwe.md) MCP server and Serena MCP: the IDE for your agent MCP server.
How does Memex compare with other AI and agents MCP servers?
safety score 66/100 (higher than 28% of 4,460 AI and agents servers); tools 13 (the typical AI and agents server has 6); tools that change things 23% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.

Alternatives to Memex

Same job from other publishers: the closest match first, then the best rated.

  • Basic Memory
    Local-first knowledge management with bi-directional LLM sync via Markdown files.
    B
  • IWE (iwe.md)
    Markdown knowledge base as agent memory. Runs against the notes directory it is started in.
    A
  • Serena MCP: the IDE for your agent
    A powerful toolkit for coding, providing semantic retrieval and editing capabilities.
    C
  • HOL Guard
    Local-first AI agent security evidence and approval workflows through HOL Guard's stdio MCP server.
    A
  • Agent-Native Clips
    Screen recording, meeting notes, and voice dictation - all with AI
    More from BuilderIO →
    A

More from evan-moon →