Voice Synthesis

TaleNode integrates with ElevenLabs to generate AI voice audio for your dialogue nodes. Assign voices to characters, then generate speech audio files that your game engine can use at runtime.

Setup

1. Get an ElevenLabs API key from elevenlabs.io
2. Open Settings > Settings...
3. Paste your API key in the ElevenLabs Voice section
4. Click Save Settings

Assigning Voices to Characters

Before generating audio, assign a voice to each character:

1. Open the Characters section in the left panel
2. Select a character to expand its editor
3. Use the Voice dropdown to pick from your ElevenLabs voice library
4. If the dropdown is empty, click Fetch Voices in the Voice Generation panel first

The dropdown shows all voices available in your ElevenLabs account, including default and custom cloned voices.

Voice Generation Panel

Open via View > Voice Generation (or drag the tab from the dock).

Panel Layout

The panel shows a table of all Dialogue nodes in your graph:

Column	Description
Speaker	Character name assigned to the dialogue node
Text	Preview of the dialogue text (truncated to 60 characters)
Voice	Assigned voice name, or "(no voice)" if the speaker's character has no voice set
Audio	Status indicator — shows the audio file path or "(none)"
Action	Generate button for individual node generation

Toolbar Buttons

Button	Description
Fetch Voices	Loads the list of available voices from your ElevenLabs account
Generate All	Generates audio for all Dialogue nodes that have an assigned voice but no audio file yet

Generating Audio

Single Node

Click the Generate button next to any dialogue node in the Voice Generation panel. The audio file is saved and the node's audio_clip field is set automatically.

Batch Generation

Click Generate All to process all dialogue nodes that:

• Have a speaker with an assigned voice
• Don't already have an audio file

Nodes are processed sequentially. Progress is shown in the status bar.

Audio Files

Generated audio files are saved to:

{project_directory}/voices/{readable_id}.mp3

For example, if your project is at /documents/my_game.talenode, audio files go to /documents/voices/dlg_1.mp3, /documents/voices/dlg_2.mp3, etc.

The audio_clip field on each Dialogue node is set to the relative path (e.g., voices/dlg_1.mp3) so your game engine can locate the files relative to the export.

Voice Model

TaleNode uses the eleven_multilingual_v2 model for generation, which supports multiple languages. This means voice generation works with translated dialogue text as well.

Tips

Tip

Fetch voices once after entering your API key. The voice list is cached for the current session — you don't need to fetch again unless you add new voices to your ElevenLabs account.

Tip

Generate audio after your dialogue text is finalized. If you change the text, you'll need to regenerate the audio for those nodes.

Tip

ElevenLabs charges by character count. Use Generate All for batch processing, but be aware of your plan's character limits. Check your usage at elevenlabs.io.

Error Handling

Error	Cause
"No ElevenLabs API key configured"	API key not set in Settings
"Character has no voice_id"	The speaker's character doesn't have a voice assigned
"ElevenLabs API error 401"	Invalid API key
"ElevenLabs API error 429"	Rate limit or quota exceeded

Check the status bar at the bottom of the window for error messages after generation attempts.