Dialogue TTS

Dialogue TTS is for game dialogue, audio drama lines, short-video scripts, and multi-character voice production.

Before Creating a Project

1. Enable at least one provider.
2. Create at least one voice character that passes Quick Test.
3. Put all required characters in the same voice bank.

Create a Project

1. Open Dialogue TTS from the left navigation.
2. Click New Project.
3. Enter a project name, such as Chapter 1 Dialogue.
4. Select a voice bank.
5. Open the project editor.

Add Lines

Use the input area at the bottom of the right panel:

1. Select a character from the voice dropdown.
2. Enter the line text.
3. Click the send button.

Each line appears as a chat bubble, which makes it easier to read, edit, and verify by character.

Check Each Line

Check	Why It Matters
Correct character	Synthesis uses the character bound to that line.
Final text	Editing text requires regenerating that line's audio.
Existing audio	Lines with audio can be played; lines without audio still need synthesis.

Generate and Play

1. Generate one line first to confirm the selected character works.
2. Click Generate All to synthesize lines without audio.
3. Failed synthesis shows an error state on the bubble.
4. Click the play button on a bubble to preview that line.
5. The waveform shows playback progress and current / total duration.

All lines use the shared TTS queue. Provider concurrency and rate limits therefore affect Dialogue TTS speed. When using cloud free quotas, generate in small batches first.

Multi-Character Script Tips

Script Shape	Recommendation
Lines already have speaker prefixes	Copy line by line and manually select the matching character.
Short-video dialogue	Use one line per shot or subtitle cue.
Audio drama	Build the full voice bank first, then review lines by character.
Same character with different emotions	Create character variants, such as Narrator Calm and Narrator Tense.

Failure Handling

Symptom	Action
One line fails	Open the error, then reproduce it in Quick TTS.
Batch fails halfway	Lower provider concurrency, keep successful lines, and retry failed lines only.
Text changed but old audio still plays	Regenerate that line.
Character dropdown misses a role	Return to Voice Bank and add the character to the project's voice bank.

Export Handling

After generation, collect audio files from the project output directory. For full-scene mixing, multi-track editing, silence spacing, or subtitle alignment, import the generated audio into an audio editor or video editor.