tts

Convert any text into speech audio. Supports two backends (Kokoro local, Noiz cloud), two modes (simple or timeline-accurate), and per-segment voice control.

Triggers

• text to speech / tts / speak / say
• voice clone / dubbing
• epub to audio / srt to audio / convert to audio
• 语音 / 说 / 讲 / 说话

Simple Mode — text to audio

speak is the default — the subcommand can be omitted:

# Basic usage (speak is implicit)
python3 skills/tts/scripts/tts.py -t "Hello world"          # add -o path to save
python3 skills/tts/scripts/tts.py -f article.txt -o out.mp3

# Voice cloning — local file path or URL
python3 skills/tts/scripts/tts.py -t "Hello" --ref-audio ./ref.wav
python3 skills/tts/scripts/tts.py -t "Hello" --ref-audio https://example.com/my_voice.wav -o clone.wav

# Voice message format
python3 skills/tts/scripts/tts.py -t "Hello" --format opus -o voice.opus
python3 skills/tts/scripts/tts.py -t "Hello" --format ogg -o voice.ogg

Third-party integration (Feishu/Telegram/Discord) is documented in ref_3rd_party.md.

Timeline Mode — SRT to time-aligned audio

For precise per-segment timing (dubbing, subtitles, video narration).

Step 1: Get or create an SRT

If the user doesn't have one, generate from text:

python3 skills/tts/scripts/tts.py to-srt -i article.txt -o article.srt
python3 skills/tts/scripts/tts.py to-srt -i article.txt -o article.srt --cps 15 --gap 500

--cps = characters per second (default 4, good for Chinese; ~15 for English). The agent can also write SRT manually.

Step 2: Create a voice map

JSON file controlling default + per-segment voice settings. segments keys support single index "3" or range "5-8".

Kokoro voice map:

{
  "default": { "voice": "zf_xiaoni", "lang": "cmn" },
  "segments": {
    "1": { "voice": "zm_yunxi" },
    "5-8": { "voice": "af_sarah", "lang": "en-us", "speed": 0.9 }
  }
}

Noiz voice map (adds emo, reference_audio support). reference_audio can be a local path or a URL (user’s own audio; Noiz only):

{
  "default": { "voice_id": "voice_123", "target_lang": "zh" },
  "segments": {
    "1": { "voice_id": "voice_host", "emo": { "Joy": 0.6 } },
    "2-4": { "reference_audio": "./refs/guest.wav" }
  }
}

Dynamic Reference Audio Slicing: If you are translating or dubbing a video and want each sentence to automatically use the audio from the original video at the exact same timestamp as its reference audio, use the --ref-audio-track argument instead of setting reference_audio in the map:

python3 skills/tts/scripts/tts.py render --srt input.srt --voice-map vm.json --ref-audio-track original_video.mp4 -o output.wav

See examples/ for full samples.

Step 3: Render

python3 skills/tts/scripts/tts.py render --srt input.srt --voice-map vm.json -o output.wav
python3 skills/tts/scripts/tts.py render --srt input.srt --voice-map vm.json --backend noiz --auto-emotion -o output.wav

When to Choose Which

When the user needs emotion control + voice cloning + precise duration together, Noiz is the only backend that supports all three.

Guest Mode (no API key)

When no API key is configured, tts.py automatically falls back to guest mode — a limited Noiz endpoint that requires no authentication. Guest mode only supports --voice-id, --speed, and --format; voice cloning, emotion, duration, and timeline rendering are not available.

# Guest mode (auto-detected when no API key is set)
python3 skills/tts/scripts/tts.py -t "Hello" --voice-id 883b6b7c -o hello.wav

# Explicit backend override to use kokoro instead
python3 skills/tts/scripts/tts.py -t "Hello" --backend kokoro

Available guest voices (15 built-in):

Requirements

• ffmpeg in PATH (timeline mode only)
• Get your API key at Noiz Developer, then run python3 skills/tts/scripts/tts.py config --set-api-key YOUR_KEY (guest mode works without a key but has limited features)
• Kokoro: if already installed, pass --backend kokoro to use the local backend

Noiz API authentication

Use only the base64-encoded API key as Authorization—no prefix (e.g. no APIKEY or Bearer ). Any prefix causes 401.

For backend details and full argument reference, see reference.md.