Advanced text-to-speech using the ElevenLabs API. Supports voice selection, language/gender filtering, audio file generation, and credit tracking.
Set your API key as an environment variable:
export ELEVENLABS_API_KEY=sk_...
Or pass it directly with --api-key.
python3 scripts/elevenlabs.py "Hello world" --voice Sarah --output audio.mp3
python3 scripts/elevenlabs.py --list-voices
python3 scripts/elevenlabs.py --list-voices --lang en --gender female
python3 scripts/elevenlabs.py --list-voices --json
python3 scripts/elevenlabs.py --credits
| Option | Default | Description |
|---|---|---|
| -------- | --------- | ------------- |
--voice NAME | Rachel | Voice name (partial match supported) |
--voice-id ID | — | Direct voice ID (overrides --voice) |
--output PATH | output.mp3 | Output MP3 file path |
--model ID | eleven_turbo_v2_5 | Model ID |
--stability | 0.5 | Voice stability (0–1) |
--similarity | 0.75 | Similarity boost (0–1) |
--style | 0.0 | Style exaggeration (0–1) |
--lang CODE | — | Filter voices by language (e.g. en, fr) |
--gender | — | Filter voices by gender (male/female) |
--json | — | Output as JSON |
--api-key KEY | — | API key (overrides env var) |
| Model ID | Description |
|---|---|
| ---------- | ------------- |
eleven_turbo_v2_5 | Fastest, lowest latency (free tier supported) |
eleven_multilingual_v2 | Best quality, multilingual |
eleven_flash_v2_5 | Ultra-low latency |
import sys
sys.path.insert(0, "path/to/skills/elevenlabs-pro/scripts")
from elevenlabs import list_voices, get_voice_id, text_to_speech, get_credits
api_key = "sk_..."
# TTS
voice_id = get_voice_id(api_key, "Sarah")
path = text_to_speech(api_key, "Hello!", voice_id, "out.mp3")
# Credits
info = get_credits(api_key)
print(info["characters_remaining"])
See references/voices.md for popular voices and voice parameter guidance.
共 1 个版本