voice-isolate SKILL.md skillA
voice-isolate from puntorigen/avatar-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add puntorigen/avatar-skills/voice-isolate@git:20260709.ec45941Pins git:20260709.ec45941 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- puntorigen/avatar-skills
- Path
voice-isolate/SKILL.md- Versions recorded
- 1
- Latest version
- git:20260709.ec45941, committed Jul 9, 2026
- Size
- 6,738 bytes
- Repository stars
- 6
- License
- MIT
Extracts clean voice samples from a video. Separates the vocal stem with Demucs, gets speech timecodes with faster-whisper, detects the recurring percussive "taka" SFX by its high-frequency signature in the accompaniment stem, drops the whole word under each detected SFX, and exports the remaining clean spoken phrases as individual sample clips plus a concatenated track. Use when the user gives a video and wants the narrator's voice isolated, clean voice samples for cloning/TTS, a voice-only…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Voice Isolate Given a video, produce **clean voice samples of the narrator**: the recurring percussive SFX (a short broadband "taka"/"tk" that bleeds over the voice) is detected and the word it lands on is dropped, so the remaining phrases are clean. The goal is good-quality voice samples (for cloning/TTS), not preserving every word — contaminated words are discarded on purpose. ## Pipeline 1. **ffmpeg** extracts the audio from the video. 2. **Demucs** (`htdemucs`) splits it into the voice stem and the accompaniment stem (everything that is not voice: music + SFX). 3. **faster-whisper** transcribes on the clean voice stem → segments + per-word timecodes (more accurate than transcribing the original mix). 4. **Taka detection (auto)**: the SFX is a short broadband transient whose energy spikes in the **high band (>5 kHz)**, while the voice body lives below ~4 kHz. In the accompaniment stem (voice already removed) those transients stand out. The HF energy envelope is peak-picked with an auto-relative threshold, so only prominent taka are flagged — not the near-silent floor. 5. **Word drop**: each detected taka is expanded to the **whole word(s)** it …
Read the whole file at its exact version.
How to install
mdr add puntorigen/avatar-skills/voice-isolate@sha256:38787d628fa770dfThe line at the top of this page pins the label git:20260709.ec45941, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_3pfik3znvfky7h4u)
4 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Jul 9, 2026.
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (6738 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
puntorigen/avatar-skills · 6 stars · license MIT · pushed 2026-09-15 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_3pfik3znvfky7h4u GET https://markdownregistry.com/api/v1/resolve?ref=puntorigen/avatar-skills/voice-isolate GET https://markdownregistry.com/api/v1/blob/38787d628fa770dfa0964a248f72ff74a0dd735e46ee5ad7c5f26f8c63425210
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.