---
title: "CLI & MCP | Murmur — Local Voice Automation for Mac"
description: "Generate local voiceovers from Terminal or let Codex, Claude Code, and Cursor call Murmur through MCP. On-device, private, and scriptable."
canonical: "https://www.murmurtts.com/automation"
---
[Murmur](https://www.murmurtts.com/)CLI + MCP

# Your voice studio, *scriptable*.

Murmur is not limited to the app window. Generate narration from Terminal, or let your coding agent create, inspect, and organize audio — all on-device, all with the models and voices already on your Mac.

[Buy Murmur](https://murmur-licenses.tarunyadav9761.workers.dev/checkout)[Download for Mac](https://murmur-updates.tarunyadav9761.workers.dev/download/latest) One-command install MCP for Codex, Claude, Cursor Apple Silicon, offline

murmur — local automation Local$`murmur generate --input video/script.md --output video/voiceover.wav`01READING SCRIPT**video/script.md**02USING VOICE**af_bella · Kokoro**03CREATED AUDIO**video/voiceover.wav** Ready for Codex, Claude, Cursor, and scripts

Two surfaces

## One local engine, two ways to drive it

Every automation rides the same app-owned pipeline — your installed models, saved voices, and license stay in one place.

For scripts

### The murmur CLI

- Turn any text or Markdown script into WAV or M4A with one command.
- Batch a whole project from a small JSON manifest you can version in git.
- Install, select, inspect, and remove local models from scripts.

$ murmur generate --text "Scene one." --voice af_bella --output audio/01.wav Created audio/01.wav · 4.2s

For agents

### The MCP server

- Codex, Claude Code, and Cursor manage models, discover voices, and create audio as tool calls.
- Agents read your script, plan scenes, and return exact file paths.
- Runs over stdio from the signed helper inside the app — no server to run.

CodexClaude CodeCursorAny MCP client

Workflow

## From install to a versioned voiceover in four steps

The CLI contract is small on purpose: status, model management, voices, generate, batch, jobs. Learn it once and every script — or every agent — can speak.

[Read the full CLI & MCP docs](https://www.murmurtts.com/docs/automation)

01

### Enable and install once

Turn on local automation in Settings → Automation and click Install CLI. Murmur links the signed helper to ~/.local/bin/murmur.

02

### Generate from Terminal

Pass text directly or point at a script file. Pick a voice and speed, choose WAV or M4A, and Murmur renders on-device.

03

### Connect your coding agent

Add one MCP server entry to Codex, Claude Code, or Cursor. The agent gains ten Murmur tools for models, voices, generation, and jobs.

04

### Batch with a manifest

Keep a scenes.json beside your video script. Relative paths resolve from the manifest folder, so the whole voiceover versions with the project.

Agent workflow

## Ask for the outcome, not the plumbing

With MCP connected, your agent can install and select a model, inspect its voices, queue a batch, poll the job, and hand back the exact files it created — while your script never leaves the workspace.

Example prompt

Read `video/script.md`. Split it into intro, demo, and outro. Use Murmur to create WAV files in `video/voiceover/`, keeping one voice and pace across every scene. Don’t replace existing takes.

01**video/voiceover/01-intro.wav**02**video/voiceover/02-demo.wav**03**video/voiceover/03-outro.wav**

MCP tools

## Ten tools your agent can call

Manage local models, discover voices, queue audio, and track or cancel jobs without opening a cloud dashboard.

`murmur_status`Setup, license, and automation readiness.`murmur_list_models`Every model, capability, and install state.`murmur_list_voices`Preset voices and saved reference profiles.`murmur_install_model`Download a model as a trackable job.`murmur_set_default_model`Choose the installed default model.`murmur_remove_model`Remove downloaded model files with confirmation.`murmur_generate_speech`Queue one WAV or M4A inside the workspace.`murmur_generate_batch`Queue a sequence of named files as one job.`murmur_get_job`Progress and finished artifact paths.`murmur_cancel_job`Stop any active long-running job.

Local by design

## Automation with boundaries

The same privacy story as the app, enforced at the helper level — so a script or an agent can only do what you would do yourself.

**Opt-in access**

Automation starts disabled. Nothing can call the helper until you allow it inside Murmur.

**One model owner**

The Mac app runs every generation, so the CLI and agents never start competing model servers.

**Workspace boundary**

MCP reads and writes only inside the directory where the agent launched it.

**No surprise overwrite**

Existing output fails safely unless you pass --force on purpose.

**Explicit model deletion**

Removing downloaded model files requires confirmation, and bundled Kokoro cannot be removed.

**No cloud hop**

Prompts, scripts, saved voices, and generated audio remain on your Mac.

## Ready to wire it up?

Install commands, agent configs for Codex, Claude Code, and Cursor, batch manifests, and the full command reference.

[Open the docs](https://www.murmurtts.com/docs/automation)

## Give every agent a local voice.

Buy once, install the CLI, and let Terminal, Codex, Claude, and Cursor narrate with the voices on your Mac.

[Buy Murmur](https://murmur-licenses.tarunyadav9761.workers.dev/checkout)[Download for Mac](https://murmur-updates.tarunyadav9761.workers.dev/download/latest)
