Skip to content
Back to search

digituva

MarketplaceWorks with Claude CodeGitHub

Claude Code plugins by Digituva Yazılım

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/digituvacom--claude-listen-audio/badge.svg)](/listings/digituvacom--claude-listen-audio)

HTML

<a href="/listings/digituvacom--claude-listen-audio"><img src="/api/listings/digituvacom--claude-listen-audio/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.

As a plugin (recommended)

Inside Claude Code, run:

/plugin install listen-audio --marketplace digituvacom/claude-listen-audio

<sub>On Claude Code older than 2.1.275, use two steps: `/plugin marketplace add digituvacom/claude-listen-audio`, then `/plugin install listen-audio@digituva`.</sub>

Then install the Python dependencies once (Python 3.9+):

python -m pip install faster-whisper librosa matplotlib soundfile

Optional, for sound-event tagging:

python -m pip install transformers torch

Updates arrive with `/plugin update listen-audio`. If a dependency is missing, Claude notices on first use and tells you what to install.

As a plain skill

git clone https://github.com/digituvacom/claude-listen-audio ~/.claude/skills/listen-audio
python -m pip install -r ~/.claude/skills/listen-audio/requirements.txt

<sub>Windows: `%USERPROFILE%\.claude\skills\listen-audio`. Restart Claude Code afterwards.</sub>

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 · 17 lines
  1. {
  2. "name": "digituva",
  3. "description": "Claude Code plugins by Digituva Yazılım",
  4. "owner": {
  5. "name": "Digituva Yazılım",
  6. "url": "https://github.com/digituvacom"
  7. },
  8. "plugins": [
  9. {
  10. "name": "listen-audio",
  11. "source": "./",
  12. "description": "Give Claude ears: Whisper transcription, acoustic analysis, sound-event tags and spectrograms for any audio or video file.",
  13. "category": "productivity",
  14. "tags": ["audio", "speech-to-text", "whisper", "spectrogram"]
  15. }
  16. ]
  17. }
README.md · README · 232 lines
  1. <div align="center">
  2. # 🎧 claude-listen-audio
  3. ### Give Claude ears.
  4. A [Claude Code](https://claude.com/claude-code) skill that turns any recording into something Claude can understand:<br>
  5. a timestamped transcript, precise acoustic measurements, and a spectrogram it can actually *look at*.
  6. [![Claude Code Skill](https://img.shields.io/badge/Claude_Code-Skill-D97757?style=for-the-badge)](https://claude.com/claude-code)
  7. [![Python 3.9+](https://img.shields.io/badge/Python-3.9+-3776AB?style=for-the-badge&logo=python&logoColor=white)](https://www.python.org/)
  8. [![Whisper](https://img.shields.io/badge/Speech-faster--whisper-412991?style=for-the-badge)](https://github.com/SYSTRAN/faster-whisper)
  9. [![License: MIT](https://img.shields.io/badge/License-MIT-22C55E?style=for-the-badge)](LICENSE)
  10. **English** · [Türkçe](README.tr.md)
  11. <a href="https://www.youtube.com/watch?v=knZadfUNXOc"><img src="https://img.youtube.com/vi/knZadfUNXOc/maxresdefault.jpg" alt="Watch the 34-second listen-audio demo on YouTube" width="860"></a>
  12. <sub>▶ <a href="https://www.youtube.com/watch?v=knZadfUNXOc">Watch the 34-second demo</a> · <a href="https://www.youtube.com/watch?v=XZR8gpV7_Ys">Türkçe video</a></sub>
  13. <br><br>
  14. <img src="docs/example-spectrogram.png" alt="Waveform, loudness curve and mel spectrogram of a test clip: speech, silence, a 440 Hz tone and white noise" width="860">
  15. <sub>One clip, four sounds: speech (0–5.7 s) · silence · a pure 440 Hz tone · white noise. Claude reads this image and tells them apart.</sub>
  16. </div>
  17. ---
  18. ## Contents
  19. - [Why](#why)
  20. - [What Claude gets](#what-claude-gets)
  21. - [How it works](#how-it-works)
  22. - [Install](#install)
  23. - [Usage](#usage)
  24. - [Example output](#example-output)
  25. - [Performance](#performance)
  26. - [Limitations](#limitations)
  27. - [FAQ](#faq)
  28. - [License](#license)
  29. ## Why
  30. Claude reads text and sees images, but it **cannot hear**. Hand it a voice note, a meeting recording or a song and it has nothing to work with.
  31. This skill closes that gap. It translates sound into the two forms Claude handles best, **numbers** and **pictures**, so it can answer questions like:
  32. > *"What does this voice note say?"* · *"Summarize this 40-minute meeting."* · *"Is there background noise or hum?"*<br>
  33. > *"What's the tempo and key of this track?"* · *"When does the music start?"* · *"Is the audio clipping?"*
  34. ## What Claude gets
  35. | | Capability | Details |
  36. |:-:|---|---|
  37. | 🗣️ | **Transcript** | Timestamped speech-to-text via faster-whisper. Auto-detects 99 languages, Turkish included. Low-confidence lines are flagged, not passed off as fact. |
  38. | 📊 | **Acoustic report** | Loudness (peak / RMS / dynamic range), clipping, silences and pauses, pitch and musical note, tempo (BPM), key, timbre, energy per frequency band. |
  39. | 🕒 | **Timeline** | How the sound evolves window by window: level, brightness, noisiness, onsets, dominant pitch class. |
  40. | 🖼️ | **Spectrogram** | Waveform + loudness curve + mel spectrogram as one PNG. Claude opens it and *sees* speech harmonics, tones, beats and noise. |
  41. | 🏷️ | **Sound events** *(optional)* | AudioSet labels per 10 s chunk: Dog, Siren, Applause, Guitar, Car horn… 527 classes. |
  42. Works with **mp3, wav, m4a, ogg, flac, opus, webm, mp4** and most other audio/video containers. No system ffmpeg required.
  43. ## How it works
  44. ```mermaid
  45. flowchart LR
  46. A[🎵 Audio / video file] --> B[Decode<br/>PyAV]
  47. B --> C[faster-whisper]
  48. B --> D[librosa analysis]
  49. B --> E[AudioSet classifier<br/><i>optional</i>]
  50. C --> F[📝 transcript]
  51. D --> G[📊 metrics + timeline]
  52. D --> H[🖼️ spectrogram.png]
  53. E --> I[🏷️ sound tags]
  54. F & G & H & I --> J((Claude))
  55. ```
  56. Everything runs **locally** on your machine. No audio is uploaded anywhere; only the model weights are downloaded once from Hugging Face.
  57. ## Install
  58. ### As a plugin (recommended)
  59. Inside Claude Code, run:
  60. ```
  61. /plugin install listen-audio --marketplace digituvacom/claude-listen-audio
  62. ```
  63. <sub>On Claude Code older than 2.1.275, use two steps: `/plugin marketplace add digituvacom/claude-listen-audio`, then `/plugin install listen-audio@digituva`.</sub>
  64. Then install the Python dependencies once (Python 3.9+):
  65. ```bash
  66. python -m pip install faster-whisper librosa matplotlib soundfile
  67. ```
  68. Optional, for sound-event tagging:
  69. ```bash
  70. python -m pip install transformers torch
  71. ```
  72. Updates arrive with `/plugin update listen-audio`. If a dependency is missing, Claude notices on first use and tells you what to install.
  73. ### As a plain skill
  74. ```bash
  75. git clone https://github.com/digituvacom/claude-listen-audio ~/.claude/skills/listen-audio
  76. python -m pip install -r ~/.claude/skills/listen-audio/requirements.txt
  77. ```
  78. <sub>Windows: `%USERPROFILE%\.claude\skills\listen-audio`. Restart Claude Code afterwards.</sub>
  79. ## Usage
  80. ### With Claude
  81. Just mention a file. The skill triggers on its own:
  82. ```
  83. > What does this say? ~/Downloads/voice-note.m4a
  84. > Summarize the decisions in meeting.mp3
  85. > Is there a 50 Hz hum in this recording?
  86. > What's the BPM and key of track.wav?
  87. ```
  88. ### From the command line
  89. ```bash
  90. python scripts/listen.py recording.mp3 --out ./out
  91. ```
  92. | Flag | Default | Description |
  93. |---|---|---|
  94. | `--mode` | `all` | `transcribe`, `analyze` or `all` |
  95. | `--model` | `small` | Whisper size: `tiny` · `base` · `small` · `medium` · `large-v3` · `turbo` |
  96. | `--language` | auto | Force a language code, e.g. `en`, `tr`, `de` |
  97. | `--tags` | off | Add AudioSet sound-event labels |
  98. | `--start` / `--end` | full file | Analyse only a time range (seconds) |
  99. | `--device` | `auto` | `cuda` or `cpu`; falls back to CPU automatically |
  100. | `--timeline` | `12` | Number of timeline windows |
  101. **Outputs** written to `--out`:
  102. ```
  103. out/
  104. ├── report.md # human-readable report (also printed to stdout)
  105. ├── report.json # every number, machine-readable
  106. ├── transcript.txt # [mm:ss → mm:ss] lines
  107. └── spectrogram.png # the picture Claude looks at
  108. ```
  109. ## Example output
  110. <details open>
  111. <summary><b>report.md</b> for the test clip shown above</summary>
  112. ```markdown
  113. ## Speech
  114. language **en** (p=0.996), 5.8 s of speech, 1 segments, 0 low-confidence
  115. [00:00.00 → 00:05.76] Hello, this is an audio test. The meeting will start at 3 o'clock this afternoon.
  116. ## Acoustics
  117. - Loudness: peak 0.0 dBFS, RMS -15.2 dBFS, dynamic range 21.3 dB, clipping 0.031%
  118. - Silence: 35% of the time, 5 sound regions; gaps ≥1s: 00:05.67–00:08.31
  119. - Pitch: voiced 50%, median 194.9 Hz (G3), range D3–A4
  120. ### Timeline
  121. | from | to | RMS dBFS | centroid Hz | flatness | pitch class |
  122. |----------|----------|----------|-------------|----------|-------------|
  123. | 00:08.24 | 00:09.27 | -11.5 | 442 | 0.0 | A | ← the 440 Hz tone
  124. | 00:11.33 | 00:12.36 | -9.8 | 5667 | 0.548 | D# | ← white noise
  125. ```
  126. </details>
  127. ## Performance
  128. Rough transcription speed with the default `small` model:
  129. | Hardware | Speed |
  130. |---|---|
  131. | NVIDIA GPU (CUDA 12) | ~20–40× real time |
  132. | Modern laptop CPU | ~1–2× real time |
  133. First use of each model downloads it once: `base` ≈ 150 MB · `small` ≈ 500 MB · `turbo` ≈ 1.6 GB.
  134. The optional tagger downloads ≈ 350 MB.
  135. ## Limitations
  136. - **No speaker diarization.** Whisper doesn't separate voices; "who said what" is inferred from content and pauses and labelled as such.
  137. - **Music metrics on speech are noise.** Tempo and key are only meaningful for music.
  138. - **Voice hints are hints.** Gender/age guesses come from median pitch alone and can be wrong.
  139. - **Pitch analysis covers the first 90 s** to keep long files fast.
  140. ## FAQ
  141. <details>
  142. <summary><b>Does my audio leave my computer?</b></summary>
  143. No. Decoding, transcription and analysis all run locally. The only network traffic is the one-time model download from Hugging Face.
  144. </details>
  145. <details>
  146. <summary><b>The GPU isn't used.</b></summary>
  147. faster-whisper needs the CUDA 12 cuBLAS and cuDNN libraries. Without them the script prints a notice and continues on CPU, so nothing breaks.
  148. </details>
  149. <details>
  150. <summary><b>The transcript is wrong in places.</b></summary>
  151. Use a larger model (`--model turbo` or `large-v3`) and, for short or mixed-language clips, force the language with `--language`.
  152. </details>
  153. <details>
  154. <summary><b>It's a 2-hour file.</b></summary>
  155. Use `--start` / `--end` for a first pass, or run the whole thing in the background. Claude will do this automatically for long files.
  156. </details>
  157. ## License
  158. [MIT](LICENSE) © 2026 Digituva Yazılım
  159. <div align="center">
  160. <sub>Built with Digituva Yazılım · Powered by <a href="https://github.com/SYSTRAN/faster-whisper">faster-whisper</a>, <a href="https://librosa.org">librosa</a> and <a href="https://matplotlib.org">matplotlib</a></sub>
  161. </div>

Ratings and comments

Sign in to rate this skill.

Sign in

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