pickfit

BrowseMCP servers

dreamd

Local-first long-term memory for AI agents with auditable lessons and provenance.

Tested on macOS and LinuxKnown vulnerabilities

Use it

Before you install: known vulnerabilities

pickfit found 2 known high or critical vulnerabilities in the packages a fresh install brings (OSV.dev, 3 Nov 2026). Weigh this before you install it.

  • crates:rmcp 1.7.0 RMCP: Missing Resource Field Validation in OAuth Protected Resource Metadata Discovery (GHSA-33f5-2c5q-wgwj, high, fixed in 2.0.0)
See 1 more

macOS arm64

  1. install dreamd-mcp globally via npm

    npm install -g dreamd-mcp
Where the README and reality differ
  • README: npx -y dreamd-mcp setup works in any folder
    What worked: requires project root sentinel (.git/, Cargo.toml, package.json or pyproject.toml); created git repo with git init

Check that it works

dreamd-mcp doctor 2>&1

The output should contain “index_freshness:”.

First thing to try

pwd; cat package.json 2>&1 || echo no-package; git init 2>&1; ls -la
dreamd-mcp setup --yes --harness none 2>&1
dreamd-mcp doctor 2>&1

Add it to your agent

claude mcp add dreamd -- dreamd-mcp
JSON for Claude Desktop, Cursor and other clients
{
  "mcpServers": {
    "dreamd": {
      "command": "dreamd-mcp",
      "args": []
    }
  }
}

Run dreamd-mcp setup --yes --harness none to set up the memory store, then add dreamd-mcp as an MCP server in your AI client.

Linux x86_64

  1. install dreamd-mcp globally via npm

    npm install -g dreamd-mcp

Check that it works

which dreamd-mcp; ls ~/.local/bin/dreamd* 2>&1; dreamd-mcp --help 2>&1

The output should contain “dreamd-mcp — npx shim”.

First thing to try

cd /tmp/testproj && dreamd-mcp init 2>&1; echo "EXIT:$?"; ls -R .agent 2>&1 | head -n 100
cd /tmp/testproj && dreamd-mcp doctor 2>&1

Add it to your agent

claude mcp add dreamd -- dreamd-mcp
JSON for Claude Desktop, Cursor and other clients
{
  "mcpServers": {
    "dreamd": {
      "command": "dreamd-mcp",
      "args": []
    }
  }
}

Run dreamd-mcp setup --yes --harness none to set up the memory store, then add dreamd-mcp as an MCP server in your AI client.

What it does

  • Search stored memories through MCP
    How to use it
    1. Run dreamd-mcp setup --yes --harness none once.
    2. Connect your MCP client to dreamd-mcp.
    3. Call the search_nodes tool with a query such as <your text>.
    Tested
  • Add new learnings through a single-writer daemon
    How to use it
    1. Run dreamd-mcp setup --yes --harness none once.
    2. Connect your MCP client to dreamd-mcp.
    3. Call the append_node tool with content, source_harness and skill_action set to <your text>.
    Tested
  • Share one memory folder across agent harnesses
    Not tested
  • Combine learnings into LESSONS.md
    Not tested

How far pickfit got

macOS arm64

  1. Installs
  2. Connects
  3. 2 of 4 features work
  4. Works offline

Linux x86_64

  1. Installs
  2. Connects
  3. 2 of 4 features work
  4. Works offline

Windows: the project says it runs there; pickfit hasn’t tried it.

Good for

  • find which stored memories drove a recall
  • remove a bad memory event and its lesson state
  • verify provenance ledger consistency
  • search prior agent context across sessions

How pickfit tested it

Each run starts from a clean machine and follows the steps above. How pickfit verifies

macOS arm64

An MCP client called one of its tools

  • Replayed from scratch
  • Works offline
  • Checked on 1 Nov 2026
  • Version 1.2.0

An MCP client handshake over stdio listed 2 tools, and a call to append_node returned a result.

Differences the test machine hit

  • README: npx -y dreamd-mcp setup with interactive harness prompt
    In the test: in non-interactive shell use --yes --harness none to avoid TTY prompt

Linux x86_64

An MCP client called one of its tools

  • Replayed from scratch
  • Works offline
  • Checked on 1 Nov 2026
  • Version 1.2.0

An MCP client handshake over stdio listed 2 tools, and a call to append_node returned a result.

Differences the test machine hit

  • README: npx -y dreamd-mcp init --yes and npx -y dreamd-mcp setup require sentinel .git/Cargo.toml/package.json/pyproject.toml
    In the test: Must create sentinel project directory with git init before handshake/tools work reliably; otherwise server cannot locate project root
  • README: npx -y dreamd-mcp as launch command
    In the test: Installed globally with npm install -g dreamd-mcp so launch is `dreamd-mcp` on PATH, avoiding npx fetch on each probe and satisfying offline replay

How to read a tool card

Stitching
pickfit installed this tool itself. Everything under the stitches comes from that run. A tool pickfit has not run never gets stitches.
Verified on your OS
The install steps ran on the system you chose. “Ran on Linux only” means it has not been run on yours yet.
macOS, Linux and WindowsWhere it ran
Green OS marks: pickfit ran the install steps there. Grey ones: the README says it runs there, and pickfit has not tried.
How far the run got
Installs and starts, its main function works, its service answers, it produces model output. A window means a desktop app opened; a plug, an MCP client used it. The mark sits beside the run’s result, under “Why it fits and how we know” and in “How pickfit tested it”.
Visual check
pickfit looked at screenshots of the app. Green: it looks right. Amber with a number: things that look wrong; open it to see them. Grey and dashed: not checked yet.
Known vulnerabilities
Known high or critical vulnerabilities (OSV.dev) in what it installs. Green: none known. Red with a number: how many; open it for the list, and read it before you install. Grey and dashed: not checked, which does not mean safe.
About the run
The day it ran, and the version that ran.
Source of a reason
Every reason a pick fits rests on one profile field. Open the reason to see the field and the value it relies on.
Hidden gem
High quality, little known.
Quality and popularity
Two scores out of 100. Quality: how well it installed and worked when pickfit tried it, and how good its install docs and upkeep are. Popularity: from stars, downloads and Hacker News.
Features checked
Of the main features the README advertises, how many the run actually exercised (green), and how many failed (red). Grey ones were not checked, and the card says why.
Not verified
A well-known tool Fitz mentions from general knowledge. No run record, no steps, no score.