Generate expressive multi-speaker audio conversations from a dialogue script using ElevenLabs V3. Perfect for podcasts, audiobooks, and interactive voice applications.
Generate expressive, multilingual text-to-dialogue content using the ElevenLabs Text To Dialogue V3 model.
About this model
ElevenLabs Text to Dialogue V3 generates natural, expressive multi-speaker conversations from a structured dialogue script. Each speaker can be assigned a distinct voice, producing studio-quality audio ideal for podcasts, audiobooks, interactive fiction, and conversational AI applications. The model supports multiple languages and provides stability controls so you can tune the consistency of each voice across longer sessions.
Built on ElevenLabs' latest V3 synthesis engine, Text to Dialogue V3 produces lifelike intonation and emotional range without post-processing. Simply define your speakers, assign voice IDs, and submit — the model handles turn-taking, pacing, and prosody automatically.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $0.10 per generation | muapiapp provides ElevenLabs Text to Dialogue V3 at a competitive flat rate with no per-character charges, making it predictable and cost-effective for multi-turn dialogues. |
| ElevenLabs (direct) | Character-based pricing (varies) | ElevenLabs charges per character consumed. Longer dialogues can quickly exceed $0.10, making muapiapp the more economical choice for multi-speaker scripts. |
muapiapp provides ElevenLabs Text to Dialogue V3 at a competitive flat rate with no per-character charges, making it predictable and cost-effective for multi-turn dialogues.
ElevenLabs charges per character consumed. Longer dialogues can quickly exceed $0.10, making muapiapp the more economical choice for multi-speaker scripts.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Dialogue Script | array | List of speaker turns. | - |
| Stability | number | Determines voice stability and randomness (0 to 1, default 0.5). | 0.5 |
| Language Code | Enum (74 options) | Target language for dialogue. Leave empty for automatic language detection. | null |
List of speaker turns.
-Determines voice stability and randomness (0 to 1, default 0.5).
0.5Target language for dialogue. Leave empty for automatic language detection.
nullDeveloper documentation
Build Your Dialogue Script
dialogue array: each item is an object with a text field (what the speaker says) and a voice_id field (the ElevenLabs voice ID for that speaker).Tune Stability (Optional)
stability (0.0–1.0, default 0.5) to control voice consistency. Higher values keep the voice steady; lower values allow more expressive variation.Set Language (Optional)
language_code (e.g. en, fr, de) to hint the target language and improve pronunciation.Submit the Request
api/v1/elevenlabs-text-to-dialogue-v3.Retrieve the Audio
Frequently asked
Each item in the `dialogue` array has a `voice_id` field. Set a different voice ID per speaker to give each character a distinct voice. You can pick from the built-in voice list or supply your own ElevenLabs custom voice ID.
The model supports a wide range of languages including English, French, German, Spanish, Portuguese, Italian, Polish, and more. Pass the appropriate `language_code` (e.g. `en`, `fr`) to improve accuracy and pronunciation for non-English scripts.
Stability (0.0–1.0) governs how consistent each voice sounds across the dialogue. A value near 1.0 keeps the voice highly uniform and predictable, while lower values introduce more expressive variation in pitch and delivery.
There is no hard cap on the number of speaker turns. You can mix as many speakers as needed within a single request by assigning different voice IDs to different dialogue items.