Mode d’emploi
- Copiez la commande et examinez le paquet avant de l’installer dans votre environnement agent.
- Consultez le document original pour les outils, identifiants et environnements requis.
- Décrivez les entrées et le livrable souhaité ; comparez la sortie aux sources.
npx skills add https://github.com/modbender/skill-library-mcp --skill edge-tts-unlimitedAvant de commencer
- L’installation n’inclut ni abonnement au modèle, ni quota API, ni accès aux fichiers.
- Étoiles et installations reflètent l’activité, pas un test indépendant de qualité ou de sécurité.
Document Skill original
Rendu depuis le dépôt source en conservant titres, exemples, code, tableaux, liens et images.
Edge TTS Unlimited
Free, unlimited neural TTS. No API key. No credits. No character limits.
Why This Over Other TTS Skills?
| Problem | Other skills | This skill |
|---|---|---|
| Long text (>5K chars) | Fail or need chunking | Works natively via --file |
| Cost | ElevenLabs burns $5/mo in 4 briefs | Free forever |
| API keys | Required | None needed |
| Server deployment | Some need native deps | Pure Python, works on Fly.io/Docker |
| 55-minute audio | Not tested/supported | Tested and confirmed |
Quick Start
Generate speech from text:
bash
scripts/speak.sh "Hello world" -o output.mp3Generate from file (recommended for long text):
bash
scripts/speak.sh --file /tmp/my-script.txt -o output.mp3With voice and speed:
bash
scripts/speak.sh --file script.txt -v en-US-GuyNeural -r "+5%" -o brief.mp3Requirements
- Python 3.8+ (pre-installed on most systems)
uvpackage manager ORpip— script auto-detects
The script auto-installs edge-tts on first run via uv run --with edge-tts (no venv needed).
Voice Presets
Use --preset for quick voice selection:
| Preset | Voice | Style |
|---|---|---|
news-us | en-US-GuyNeural +5% | Lively US news anchor |
news-bbc | en-GB-RyanNeural | Authoritative British |
calm | en-US-AndrewNeural -10% | Warm, relaxed |
fast | en-US-ChristopherNeural +20% | Speed-read |
bash
scripts/speak.sh --file brief.txt --preset news-us -o brief.mp3All Options
scripts/speak.sh [TEXT] [OPTIONS]
TEXT Text to speak (or use --file)
--file, -f FILE Read text from file (recommended for long content)
--voice, -v NAME Voice name (default: en-US-GuyNeural)
--rate, -r RATE Speed adjustment: "+5%", "-10%", etc.
--preset, -p NAME Use a voice preset (see above)
--output, -o FILE Output path (default: /tmp/tts-{timestamp}.mp3)
--list List available voices
--list-filter STR Filter voice list (e.g. "british", "female")Popular Voices
Run scripts/speak.sh --list for all voices, or filter:
bash
scripts/speak.sh --list-filter british
scripts/speak.sh --list-filter femaleEnglish highlights:
en-US-GuyNeural— Male, news/passion (best for briefs)en-US-ChristopherNeural— Male, authoritativeen-US-AriaNeural— Female, confidenten-GB-RyanNeural— Male, British steadyen-GB-SoniaNeural— Female, British
Tested Limits
| Chars | Duration | File Size | Result |
|---|---|---|---|
| 7,400 | ~6 min | 2.6 MB | ✅ |
| 18,000 | ~21 min | 7.3 MB | ✅ |
| 46,200 | ~55 min | 18.8 MB | ✅ |
No hard limit found. Microsoft Edge TTS service handles arbitrarily long text.
Tips
- Always use `--file` for text longer than a sentence — avoids shell quoting issues
- Rate "+5%" sounds natural for news; "+20%" for speed-listening
- Edge TTS interprets
[short pause]as literal text — use commas and periods for natural pauses - Audio is 48kbps mono MP3 by default — good enough for voice, small file size