Earshot MCP Server

Analyzes local audio, checks timing against video, and exposes sound measurements and synthesis tools through MCP.

Published by LeonxlnxLeonxlnx/earshot

.install .install-tabs{display:none}.install .install-panel[hidden]{display:block}@media (max-width:767px){.install .install-panel[hidden]{display:flex}}.install .install-panel+.install-panel{border-top:1px solid var(--line)}.install .install-panel-label{display:block}.install .install-all{display:none}

Install

Pick your client, then copy the command.

Claude Code

Run in your terminal

claude mcp add earshot -- /path/to/earshot/.venv/bin/python -m earshot mcp

Codex

~/.codex/config.toml — generic MCP template; upstream does not name Codex. Required credentials must be present in the local Codex environment.

[mcp_servers.earshot]
command = "/path/to/earshot/.venv/bin/python"
args = ["-m", "earshot", "mcp"]

Cursor

.cursor/mcp.json in a project, or ~/.cursor/mcp.json

{
  "mcpServers": {
    "earshot": {
      "command": "/path/to/earshot/.venv/bin/python",
      "args": [
        "-m",
        "earshot",
        "mcp"
      ]
    }
  }
}

Setup

Run locally after reviewing the pinned instructions

(
set -e
git clone https://github.com/Leonxlnx/earshot.git
cd earshot && git checkout 925ebe56a1b755e5d354ab642bb8f344440cfa9c
python -m venv .venv
.venv/bin/python -m pip install torch --index-url https://download.pytorch.org/whl/cpu
.venv/bin/python -m pip install -e ".[mcp]"
.venv/bin/python -m earshot doctor
)

Claude Desktop

claude_desktop_config.json

{
  "mcpServers": {
    "earshot": {
      "command": "/path/to/earshot/.venv/bin/python",
      "args": [
        "-m",
        "earshot",
        "mcp"
      ]
    }
  }
}

Git

Check out the pinned revision

git clone 'https://github.com/Leonxlnx/earshot.git'
cd earshot && git checkout 925ebe56a1b755e5d354ab642bb8f344440cfa9c

Any agent

Paste into your agent chat

Read the source at https://github.com/Leonxlnx/earshot/blob/925ebe56a1b755e5d354ab642bb8f344440cfa9c/README.md (pinned revision 925ebe56a1b7). Explain what it does and which tools, credentials and network access it needs. Do not install or run anything until I confirm.

All client instructions (works without JavaScript)

Claude Code

Run in your terminal

claude mcp add earshot -- /path/to/earshot/.venv/bin/python -m earshot mcp

Codex

~/.codex/config.toml — generic MCP template; upstream does not name Codex. Required credentials must be present in the local Codex environment.

[mcp_servers.earshot]
command = "/path/to/earshot/.venv/bin/python"
args = ["-m", "earshot", "mcp"]

Cursor

.cursor/mcp.json in a project, or ~/.cursor/mcp.json

{
  "mcpServers": {
    "earshot": {
      "command": "/path/to/earshot/.venv/bin/python",
      "args": [
        "-m",
        "earshot",
        "mcp"
      ]
    }
  }
}

Setup

Run locally after reviewing the pinned instructions

(
set -e
git clone https://github.com/Leonxlnx/earshot.git
cd earshot && git checkout 925ebe56a1b755e5d354ab642bb8f344440cfa9c
python -m venv .venv
.venv/bin/python -m pip install torch --index-url https://download.pytorch.org/whl/cpu
.venv/bin/python -m pip install -e ".[mcp]"
.venv/bin/python -m earshot doctor
)

Claude Desktop

claude_desktop_config.json

{
  "mcpServers": {
    "earshot": {
      "command": "/path/to/earshot/.venv/bin/python",
      "args": [
        "-m",
        "earshot",
        "mcp"
      ]
    }
  }
}

Git

Check out the pinned revision

git clone 'https://github.com/Leonxlnx/earshot.git'
cd earshot && git checkout 925ebe56a1b755e5d354ab642bb8f344440cfa9c

Any agent

Paste into your agent chat

Read the source at https://github.com/Leonxlnx/earshot/blob/925ebe56a1b755e5d354ab642bb8f344440cfa9c/README.md (pinned revision 925ebe56a1b7). Explain what it does and which tools, credentials and network access it needs. Do not install or run anything until I confirm.

Listed ≠ audited. Snippets follow the pinned source docs; review the source and any scripts before running them. Nothing is installed from this page.

Source record

Repository: Leonxlnx/earshot
Source file: README.md
Pinned revision: 925ebe5
Upstream commit: 18 Oct 2026
Last checked: 19 Oct 2026
Stars at check: 816 (2026-10-19)
Repository created: 2026-10-15
Default branch: main
License: MIT
Cited in guides: 1

No review scores. No automatic installs. Not affiliated with the publisher.

What you are opening

Use it to inspect a sound file or audit a video soundtrack with measurements, tags and visual evidence. The sound bank can render effects without a model; semantic listening loads local models.

How to set it up

Create a virtual environment in the pinned checkout and install CPU PyTorch and the MCP extra.
Use the absolute virtual-environment Python path in the MCP tab; replace /path/to/earshot first.
Run earshot doctor, then try a measurement on a file you own.

Requirements and third-party terms

Python 3.10+, a virtual environment and CPU PyTorch. First model use downloads about 1.1 GB: CLAP (Apache-2.0) and AST (BSD-3-Clause). Optional Gemini critique sends audio only with explicit upload consent and your own key; provider terms and charges apply. FFmpeg is needed for video workflows.

Before you add it

Read the instructions and review any scripts before running them.
Check tool access, required credentials and third-party network requests.
The source license and any resource-specific terms take precedence.

What you are opening

How to set it up

Requirements and third-party terms

Before you add it

Other MCP servers