agentleFS
AlexAnsartAlexAnsart/demo-studio/skills/narrate/SKILL.md

narrate

Skill3 starsChanged 4 months agoMITView on GitHub

Generates a narration voiceover .mp3 for a demo video script using the ElevenLabs API — configurable voice/model/stability, deterministic output filename. Use when the user asks to add a voiceover, narration, or spoken audio to a demo/tutorial video, or names ElevenLabs explicitly.

  • Reads credentials
  • Installs packages

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. tslateman/skills★ 1same name
    2. vybzai/spitfire-storybook★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository4 other files it gives its agents.

The file

Files
SKILL.md

Narrate — generate a voiceover with ElevenLabs

Turns a plain-text script into an .mp3 narration track. This is stage 1 of the full pipeline (produce-video); use it standalone if you already have a silent video and just need the voice.

Requires ELEVENLABS_API_KEY (see .env.example at the project root — get a key at https://elevenlabs.io). No key, no narration — produce-video and film-demo both work fine without this skill (silent video only).

Quick start

pip install -r scripts/requirements.txt   # once

python scripts/generate_narration.py \
  --out demos/<slug>/narration.mp3 \
  --text "Here's how it works. Type your task, hit add, and watch it appear instantly."

Or --text-file script.txt for longer scripts. The script writes exactly one .mp3 at --out — no auto-naming, so it composes cleanly into a pipeline.

Defaults (override via flags or demo.config.json → narration)

SettingDefaultFlag
Voice"Rachel" (search by name in your ElevenLabs voice library)--voice-name
Voice ID (skips the name search)—--voice-id
Modeleleven_v3--model
Stability1.0 (Robust — v3 scale: 0.0 Creative, 0.5 Natural, 1.0 Robust)--stability
Output formatmp3_44100_128--format

Pick a voice from your own ElevenLabs account (or a shared voice you've added to your library) and set it once in demo.config.json → narration.voiceName so every video in the project uses the same voice. --voice-name does a voices.search() by exact name — if you get "voice not found", the voice either isn't in your account yet (add it from the ElevenLabs Voice Library first) or the name doesn't match exactly; pass --voice-id instead to skip the search entirely.

Script-writing rules (these make sync easier later)

  • Plain sentences, one idea each. sync-narration aligns and cuts at sentence boundaries — a 40-word sentence covering three on-screen actions cannot be split apart later.
  • Write to match the pacing you'll film, not the other way around. ~150 words/minute is natural TTS speed — a 60-90s video needs roughly 150-220 words of narration.
  • No stage directions or brackets ("[pause]", "(click here)") — the model reads them aloud. Convey pacing with actual punctuation and sentence breaks.
  • Say what's on screen in the order it happens. sync-narration maps sentence order to beat order 1:1 — reordering later means re-aligning.

Gate before moving on

ffprobe -v error -show_entries format=duration -of csv=p=0 demos/<slug>/narration.mp3

Confirm the duration is plausible for the word count (~150 wpm). Regenerating costs API credits — only do it when the script text actually changes; reuse the existing .mp3 otherwise.

Deliverable

Report the output path, duration, voice/model used, and word count. If ELEVENLABS_API_KEY is missing, say so explicitly and stop — do not silently fall back to a different provider or skip narration without telling the user.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.