Skip to content
Back to search

jarrodwatts-claude-stt

MarketplaceWorks with Claude CodeGitHub

Speech-to-text input for Claude Code with live streaming dictation

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/jarrodwatts-claude-stt/badge.svg)](/listings/jarrodwatts-claude-stt)

HTML

<a href="/listings/jarrodwatts-claude-stt"><img src="/api/listings/jarrodwatts-claude-stt/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Inside a Claude Code instance, run the following commands:

**Step 1: Add the marketplace**

/plugin marketplace add jarrodwatts/claude-stt

**Step 2: Install the plugin**

/plugin install claude-stt

**Step 3: Run setup**

/claude-stt:setup

Done! Press **Ctrl+Shift+Space** to start recording, press again to stop and transcribe.

> **Note**: Setup installs dependencies (uv if available, otherwise a local `.venv`), > downloads the Moonshine model (~200MB), and checks microphone permissions.

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 31 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "jarrodwatts-claude-stt",
  3. "owner": {
  4. "name": "Jarrod Watts",
  5. "url": "https://github.com/jarrodwatts"
  6. },
  7. "plugins": [
  8. {
  9. "name": "claude-stt",
  10. "source": "./",
  11. "description": "Speech-to-text input for Claude Code with live streaming dictation. Hold a hotkey, speak, and your words appear in the input field.",
  12. "version": "0.1.0",
  13. "author": {
  14. "name": "Jarrod Watts",
  15. "url": "https://github.com/jarrodwatts"
  16. },
  17. "homepage": "https://github.com/jarrodwatts/claude-stt",
  18. "repository": "https://github.com/jarrodwatts/claude-stt",
  19. "license": "MIT",
  20. "keywords": [
  21. "speech-to-text",
  22. "stt",
  23. "voice",
  24. "dictation",
  25. "moonshine",
  26. "audio",
  27. "accessibility"
  28. ]
  29. }
  30. ]
  31. }
README.md · README · 198 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # 🚨 This project is archived
  2. Native voice/STT support has been officially added to Claude Code via the `/voice` command — use that instead. This repo is no longer maintained.
  3. ---
  4. # Claude STT
  5. Speech-to-text input for Claude Code. Hold a hotkey, speak, and your words appear in the input field — all processed locally.
  6. [![License](https://img.shields.io/github/license/jarrodwatts/claude-stt)](LICENSE)
  7. [![Stars](https://img.shields.io/github/stars/jarrodwatts/claude-stt)](https://github.com/jarrodwatts/claude-stt/stargazers)
  8. ![Claude STT in action](preview.png)
  9. ## Install
  10. Inside a Claude Code instance, run the following commands:
  11. **Step 1: Add the marketplace**
  12. ```
  13. /plugin marketplace add jarrodwatts/claude-stt
  14. ```
  15. **Step 2: Install the plugin**
  16. ```
  17. /plugin install claude-stt
  18. ```
  19. **Step 3: Run setup**
  20. ```
  21. /claude-stt:setup
  22. ```
  23. Done! Press **Ctrl+Shift+Space** to start recording, press again to stop and transcribe.
  24. > **Note**: Setup installs dependencies (uv if available, otherwise a local `.venv`),
  25. > downloads the Moonshine model (~200MB), and checks microphone permissions.
  26. ---
  27. ## What is Claude STT?
  28. Claude STT gives you voice input directly into Claude Code. No typing required — just speak naturally.
  29. | What You Get | Why It Matters |
  30. |--------------|----------------|
  31. | **Local processing** | All audio processed on-device using Moonshine STT |
  32. | **Low latency** | ~400ms transcription time |
  33. | **Push-to-talk** | Hold hotkey to record, release to transcribe |
  34. | **Cross-platform** | macOS, Linux, Windows |
  35. | **Privacy first** | No audio or text sent to external services |
  36. ### How It Works
  37. ```
  38. Press Ctrl+Shift+Space → start recording
  39. ↓
  40. Audio captured from microphone
  41. ↓
  42. Press Ctrl+Shift+Space → stop recording
  43. ↓
  44. Moonshine STT processes locally (~400ms)
  45. ↓
  46. Text inserted into Claude Code input
  47. ```
  48. **Key details:**
  49. - Audio is processed in memory and immediately discarded
  50. - Uses Moonshine ONNX for fast local inference
  51. - Keyboard injection or clipboard fallback
  52. - Native system sounds for audio feedback
  53. ---
  54. ## Configuration
  55. Customize your settings anytime:
  56. ```
  57. /claude-stt:config
  58. ```
  59. ### Options
  60. | Option | Values | Default | Description |
  61. |--------|--------|---------|-------------|
  62. | `hotkey` | Key combo | `ctrl+shift+space` | Trigger recording |
  63. | `mode` | `toggle`, `push-to-talk` | `toggle` | Press to toggle vs hold to record |
  64. | `engine` | `moonshine`, `whisper` | `moonshine` | STT engine |
  65. | `moonshine_model` | `moonshine/tiny`, `moonshine/base`, other Moonshine model IDs | `moonshine/base` | Model size |
  66. | `output_mode` | `auto`, `injection`, `clipboard` | `auto` | How text is inserted |
  67. | `sound_effects` | `true`, `false` | `true` | Play audio feedback |
  68. | `max_recording_seconds` | 1-600 | 300 | Maximum recording duration |
  69. Settings stored in `~/.claude/plugins/claude-stt/config.toml`.
  70. ---
  71. ## Requirements
  72. - **Python 3.10-3.13**
  73. - **~200MB disk space** for STT model
  74. - **Microphone access**
  75. ### Platform-Specific
  76. | Platform | Additional Requirements |
  77. |----------|------------------------|
  78. | **macOS** | Accessibility permissions (System Settings > Privacy & Security) |
  79. | **Linux** | xdotool for window management; X11 recommended (Wayland has limitations); WSL not supported |
  80. | **Windows** | pywin32 for window tracking |
  81. ---
  82. ## Commands
  83. | Command | Description |
  84. |---------|-------------|
  85. | `/claude-stt:setup` | First-time setup: check environment, install deps, download model |
  86. | `/claude-stt:start` | Start the STT daemon |
  87. | `/claude-stt:stop` | Stop the STT daemon |
  88. | `/claude-stt:status` | Show daemon status and readiness checks |
  89. | `/claude-stt:config` | Change settings |
  90. You can also use the CLI directly:
  91. ```
  92. claude-stt setup
  93. claude-stt start --background
  94. ```
  95. ---
  96. ## Troubleshooting
  97. | Issue | Solution |
  98. |-------|----------|
  99. | No audio input | Check microphone permissions in system settings |
  100. | Keyboard injection not working | **macOS**: Grant Accessibility permissions. **Linux**: Ensure xdotool installed |
  101. | Model not loading | Run `/claude-stt:setup` to download. Check disk space (~200MB) |
  102. | Hotkey test fails during setup | Fix permissions or rerun `/claude-stt:setup --skip-hotkey-test` to continue setup |
  103. | Whisper dependencies missing | Run `/claude-stt:setup --with-whisper`, or `uv sync --directory $CLAUDE_PLUGIN_ROOT --extra whisper`, or `python $CLAUDE_PLUGIN_ROOT/scripts/exec.py -m pip install .[whisper]` |
  104. | Hotkey not triggering | Check for conflicts with other apps. Try `/claude-stt:config` to change hotkey |
  105. | Text going to wrong window | Plugin tracks original window — ensure Claude Code was focused when recording started |
  106. | Running under WSL | Not supported; use native Windows or Linux |
  107. ### Logging
  108. Set `CLAUDE_STT_LOG_LEVEL=DEBUG` to get verbose logs when starting the daemon.
  109. ---
  110. ## Privacy
  111. **All processing is local:**
  112. - Audio captured from your microphone is processed entirely on-device
  113. - Moonshine runs locally — no cloud API calls
  114. - Audio is never sent anywhere, never stored (processed in memory, discarded)
  115. - Transcribed text only goes to Claude Code input or clipboard
  116. **No telemetry or analytics.**
  117. ---
  118. ## Development
  119. ```bash
  120. git clone https://github.com/jarrodwatts/claude-stt
  121. cd claude-stt
  122. # Install dependencies (uv preferred, falls back to local venv)
  123. python scripts/setup.py --dev --skip-audio-test --skip-model-download --no-start
  124. # Test locally without installing
  125. claude --plugin-dir /path/to/claude-stt
  126. ```
  127. See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
  128. ### Release Checklist
  129. - Bump versions in `pyproject.toml`, `.claude-plugin/plugin.json`, `.claude-plugin/marketplace.json`
  130. - Update `CHANGELOG.md`
  131. - Run tests: `uv run python -m unittest discover -s tests`
  132. - Verify onboarding in Claude Code (`/plugin install`, `/claude-stt:setup`)
  133. ---
  134. ## License
  135. MIT — see [LICENSE](LICENSE)
  136. ---
  137. ## Star History
  138. [![Star History Chart](https://api.star-history.com/svg?repos=jarrodwatts/claude-stt&type=Date)](https://star-history.com/#jarrodwatts/claude-stt&Date)

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.