1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
const axios = require('axios');
const api_key = "YOUR API-KEY";
const url = "https://api.segmind.com/v1/tts-eleven-labs";
const data = {
"prompt": "Unlock endless possibilities with voice synthesis, bringing text to life with realism and emotion.",
"voice": "Rachel",
"voice_id": "21m00Tcm4TlvDq8ikWAM",
"model_id": "eleven_multilingual_v2",
"stability": 0.5,
"use_speaker_boost": true,
"similarity_boost": 0.75,
"style": 0,
"speed": 1,
"seed": 0,
"apply_text_normalization": "auto",
"apply_language_text_normalization": false
};
(async function() {
try {
const response = await axios.post(url, data, { headers: { 'x-api-key': api_key } });
console.log(response.data);
} catch (error) {
console.error('Error:', error.response.data);
}
})();Attributes
The text to be converted to audio. Craft engaging content to leverage audio features.
Select the voice persona for audio. Try different voices for varied audience impact.
Allowed values:
Identifier for specific audio voice. Adjust when using custom voice models.
Model selection for TTS conversion. Use multilingual for diverse language support.
Allowed values:
Enforce language in ISO 639-1 format. Choose based on target audience linguistic needs.
Allowed values:
Control voice emotional stability. Use lower values for more variance in emotion.
min : 0,
max : 1
Boost speaker likeness for more realistic voices. Enable to match original speaker.
Adjust voice fidelity to the original. Higher values ensure closer match to the voice.
min : 0,
max : 1
Define the stylistic tone of the voice. Increase for dramatic delivery.
min : 0,
max : 1
Adjust the voice reading speed. Use 1.0 for normal pace; modify for speed variation.
min : 0.25,
max : 4
Set a seed to ensure reproducibility. Use 0 for non-deterministic output.
min : 0,
max : 4294967295
Decide on text normalization. Auto is preferred for balanced results.
Allowed values:
Specific to Japanese text normalization. Enable for correct text processing.
To keep track of your credit usage, you can inspect the response headers of each API call. The x-remaining-credits property will indicate the number of remaining credits in your account. Ensure you monitor this value to avoid any disruptions in your API usage.
ElevenLabs TTS: Text-to-Speech Model
What is ElevenLabs TTS?
ElevenLabs TTS is a family of AI text-to-speech models that converts written text into natural, expressive, human-like speech. It’s designed for developers building anything from long-form narration to real-time conversational experiences, with strong prosody, contextual delivery, and multilingual output.
On Segmind, you can select from multiple ElevenLabs model IDs depending on your latency and quality needs: eleven_v3 for rich emotional delivery and dialogue, eleven_multilingual_v2 for consistent long-form narration across many languages, and eleven_flash_v2_5 / eleven_turbo_v2_5 for low-latency TTS in interactive apps. You can also choose a preset voice (like “Rachel”) or provide a specific voice_id for tighter control.
Key Features
- High-fidelity speech synthesis with expressive intonation and natural pacing
- Multiple model options for quality vs latency (v3, multilingual v2, flash/turbo)
- Voice selection via preset voices or explicit
voice_id - Multilingual TTS with optional
language_code(ISO 639-1) - Controllable prosody using
stability,similarity_boost,style, andspeed - Reproducible outputs with
seedfor consistent generations
Best Use Cases
- Audiobooks and long-form narration (consistent tone across chapters)
- Podcast intros, ads, and video voiceovers (clean, studio-like delivery)
- Real-time conversational AI and IVR (low latency with flash/turbo)
- Games and character dialogue (distinct voices + dramatic style control)
- Accessibility and assistive reading tools (clear pacing and pronunciation)
- Social content localization (multilingual voiceovers at scale)
Prompt Tips and Output Quality
- Write as you want it spoken: short sentences, intentional punctuation, and paragraph breaks.
- For more emotional variance, lower
stability(e.g., 0.2–0.4). For steady narration, raise it (0.6–0.9). - Increase
stylefor more dramatic delivery; keep near 0 for neutral reads. - Use
similarity_boost(e.g., 0.7–0.9) to keep the voice closer to the target persona. - Adjust
speed(1.0 is natural; 0.85 for gravitas; 1.1–1.3 for energetic explainers). - If you need consistent results in testing, set a fixed
seed(use0for non-deterministic).
FAQs
Is ElevenLabs TTS open-source?
No. These are proprietary text-to-speech models exposed via API-style parameters on Segmind.
Which model_id should I choose?
Use eleven_multilingual_v2 for long-form multilingual narration, eleven_v3 for expressive performances, and eleven_flash_v2_5 / eleven_turbo_v2_5 for low-latency, real-time TTS.
How do I pick between voice and voice_id?
Use voice for quick preset selection. Use voice_id when you need a specific exact voice identity (including custom voices).
What parameters most affect realism?
Start with stability, similarity_boost, and use_speaker_boost. Then tune style and speed for delivery.
Should I set language_code?
Set language_code to force a target language (e.g., en, es, ja) when your text or app supports multiple locales.
What does text normalization do?
apply_text_normalization (auto/on/off) controls how numbers, dates, and abbreviations are expanded for speech; leave auto for most apps.
Popular Models
