TTS BigBossInstall

Summary

Convert text to speech using the selected TTS engine, outputting audio as a binary WAV property and optional SRT subtitles as another binary property.

Inputs

• engine — TTS engine to use for synthesis. Options: edge (Edge TTS online, requires internet, supports multilingual voices and perfect subtitles), piper_local (offline Piper TTS, auto-downloads models), coqui (connect to a running Coqui TTS/XTTS server), kokoro (local Kokoro model generation), kokoro_sentence (Kokoro with per-sentence chunking for exact SRT timing), mozilla (local tts CLI, requires pip install TTS), or system (custom command for installed tools like XTTS). Default: edge. Display options show per-engine settings (e.g., edgeVoice/edgeRate/pitch only when engine=edge).
• text (required) — The text to convert to speech. Empty or whitespace-only text throws a NodeOperationError.
• audioProperty — Name of the binary property to output the generated audio to. Default: 'audio'. The binary contains base64-encoded WAV audio with mimeType 'audio/wav'.
• srtProperty — Name of the binary property to output the SRT subtitles to. Default: 'subtitle'. The binary contains base64-encoded SRT text with mimeType 'application/x-subrip'.

Output shape

Each input item produces one output item with json metadata and binary properties.

The operation splits text into chunks if exceeding 300 characters (MAX_CHARS_PER_CHUNK), concatenates multiple chunks with 200ms silence gaps, and generates SRT subtitles accordingly. Supported engines handle chunking and SRT generation differently: edge uses word-boundary metadata from the Bing websocket; kokoro_sentence splits by sentence boundaries; other engines use a heuristic duration model. Audio is always returned as base64-encoded WAV format.

Examples

Example 1: Basic Edge TTS synthesis

engine=edge, text='Hello world', audioProperty='myAudio', srtProperty='mySubtitle'

Links

• n8n community nodes installation guide