Earshot is a toolkit that lets AI coding agents analyse and describe audio, check how well sound fits a description or a video, and create their own sound effects in code. It runs locally on a CPU as a command-line tool, a Python library or an MCP server.
marrow-mcp
by getmarrow
This is an MCP server that connects AI coding agents such as Claude Code, Cursor and Codex to Marrow AI, a service that checks policy and past lessons before an agent takes a significant action. It then records proof of the outcome so later agents can reuse it.
Proof level: ReadWorth a lookMIT
No demo yet
Proof level: Read. We read its README and licence and ran our automatic checks. We have not run it yet. So this page is based on its README and our automatic checks.
The catch
Check this first: Needs an account and API key for the paid Marrow AI service, so the MCP package is only a client.
Things to check
- Needs an account and API key for the paid Marrow AI service, so the MCP package is only a client
- Free and Team plans only warn rather than block actions; full enforcement needs the evaluation or Business plan
- Hooks run inside the agent host and are not an external security boundary
- Native Windows is not supported; WSL is required
- Small project with few stars
Can you use marrow-mcp commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Did it work for you?
One answer per person for each tool a day, with no account and no cookies. Your answer is remembered in this browser only. How we count
What can you do with marrow-mcp?
- Hold risky agent actions such as deploys and publishes until a policy check approves them
- Record proof and outcomes for finished agent tasks in an audit trail
- Share lessons from earlier failures so the next agent avoids repeating them
How do you get marrow-mcp running?
Install with npx (npx -y @getmarrow/install@latest activate) after creating an API key at getmarrow.ai and saving it as MARROW_API_KEY; Windows needs WSL.
Follow the official instructions in the project README.
Add it to Claude Code
claude mcp add marrow -- npx -y @getmarrow/mcp
It needs MARROW_API_KEY (a secret key). A key usually means a paid or sign-up service behind it, so check its pricing.
From the official MCP Registry (ai.getmarrow/marrow). Only connect MCP servers you trust: they can act on your behalf.
Make money with it
A freelancer could offer set-up of governed AI agent workflows with approval and audit trails for small teams.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Put approval checks on AI agents that deploy code or message customers, reducing costly mistakes
- Keep an audit trail of what agents did, to support client reporting and internal review
Checked 08 Nov 2026. Summary drafted by AI from the project's README. Proof level 1 of 4, Read: we read its readme and licence and ran our automatic checks. we have not run it yet.
Similar tools in MCP servers
Kanzo is a small kanban board that runs in your browser and also works as an MCP server, so you and your AI coding agents share the same task list. All data sits in one local folder, which can be synced between machines through a private git repository.
A collection of 77 unofficial MCP servers that let AI assistants search and read Iranian websites such as Digikala, Torob, Divar, Snappfood and Alibaba. It is for comparing prices in Toman, finding jobs, cars, travel, laws and market data from Claude, Cursor or any MCP client.