#is the elevenlabs v3 speech to text available in the API?
1 messages · Page 1 of 1 (latest)
No. There is no “v3” Speech-to-Text model in the API. The Speech-to-Text endpoint currently accepts only scribe_v1 (and scribe_v1_experimental) as model_id values. (elevenlabs.io)
“Eleven v3” refers to the latest Text-to-Speech model (speech synthesis), which is available via the TTS API using modelId=eleven_v3—not for transcription. (elevenlabs.io)
Sources:
Create transcript | ElevenLabs Documentation
Models | ElevenLabs Documentation
oh sorry I meant Text to Speech, can I use the v3 in the API, do you have an example?
Yes. Eleven v3 Text to Speech is available via the API. Use the Create speech endpoint and set model_id to eleven_v3. (elevenlabs.io)
Minimal examples:
-
cURL
curl -X POST "https://api.elevenlabs.io/v1/text-to-speech/JBFqnCBsd6RMkjVDRZzb?output_format=mp3_44100_128"
-H "xi-api-key: $XI_API_KEY"
-H "Content-Type: application/json"
-d '{ "text": "Hello from v3.", "model_id": "eleven_v3" }'
(elevenlabs.io) -
Node.js (official SDK)
import { ElevenLabsClient } from '@elevenlabs/elevenlabs-js';
const client = new ElevenLabsClient({ apiKey: process.env.ELEVENLABS_API_KEY });
const audio = await client.textToSpeech.convert('JBFqnCBsd6RMkjVDRZzb', {
text: 'Hello from v3.',
modelId: 'eleven_v3',
outputFormat: 'mp3_44100_128',
});
(elevenlabs.io) -
Python (official SDK)
from elevenlabs import ElevenLabs
client = ElevenLabs(api_key="YOUR_API_KEY")
audio = client.text_to_speech.convert(
voice_id="JBFqnCBsd6RMkjVDRZzb",
text="Hello from v3.",
model_id="eleven_v3",
output_format="mp3_44100_128",
)
(elevenlabs.io)
Notes:
- v3 is currently labeled “alpha,” optimized for expressive, long‑form generation (not real‑time/Conversational AI). It supports 70+ languages and has a 3,000‑character per‑request limit. (elevenlabs.io)
References:
- API availability announcement (August 20, 2025) and instructions to set model_id=eleven_v3. (elevenlabs.io)
- Create speech endpoint (POST /v1/text-to-speech/:voice_id). (elevenlabs.io)
- Eleven v3 model overview and SDK example. (elevenlabs.io)
Sources:
August 20, 2025 | ElevenLabs Documentation
Create speech | ElevenLabs Documentation
Eleven v3 (alpha) — The most expressive Text to Speech model
Create speech | ElevenLabs Documentation
Models | ElevenLabs Documentation
August 20, 2025 | ElevenLabs Documentation
Create speech | ElevenLabs Documentation
Models | ElevenLabs Documentation
can I use it with .textToSpeech.convertAsStream ?
Short answer: use textToSpeech.stream, not .convertAsStream.
Details:
- In the Node/TypeScript SDK, the documented method for streaming Text-to-Speech is textToSpeech.stream (maps to POST /v1/text-to-speech/:voice_id/stream). (elevenlabs.io)
- The method name convertAsStream is documented for Speech-to-Speech (voice changer), not Text-to-Speech. (elevenlabs.io)
- To use the Eleven v3 TTS model with streaming, call textToSpeech.stream and set modelId: 'eleven_v3'. The v3 model ID for TTS is eleven_v3 per the official model docs. (elevenlabs.io)
- If you need timing info with streaming, use the “Stream with timestamps” TTS endpoint. (elevenlabs.io)
So: elevenlabs.textToSpeech.stream(voiceId, { text, modelId: 'eleven_v3', ... }). (elevenlabs.io)
Sources:
Streaming | ElevenLabs Documentation
Voice changer stream
Models | ElevenLabs Documentation
Stream speech with timing | ElevenLabs Documentation
Streaming | ElevenLabs Documentation
does the model eleven_flash_v3 works?
No. “eleven_flash_v3” is not a valid model ID in the public ElevenLabs docs. The Flash models available via the API are:
- eleven_flash_v2 (English)
- eleven_flash_v2_5 (multilingual)
See the Models overview and the Flash announcement, which list only v2 and v2.5. If you’re looking for the new v3 capabilities, that’s the separate “eleven_v3” model (alpha), not a Flash model. (elevenlabs.io)
Sources:
Models | ElevenLabs Documentation
is the eleven_v3 still in beta?
No. The model ID eleven_v3 is in public alpha (research preview), not beta. It’s covered by the Beta Services Addendum, so content generated with v3 (Alpha) cannot be used commercially or in production. (help.elevenlabs.io)
when will it be available?
Short answer: It’s already available.
- Eleven v3 (Alpha) has been available in the API since August 20, 2025. Use model_id "eleven_v3" on the Text-to-Speech (and Text-to-Dialogue) endpoints. (elevenlabs.io)
- It remains a public alpha under the Beta Services Addendum, so content generated with v3 (Alpha) cannot be used commercially or in production. No general-availability date is published in the docs. (help.elevenlabs.io)
Sources:
August 20, 2025 | ElevenLabs Documentation
What is Eleven v3 (Alpha)? – ElevenLabs
can I use tags [emotion] in the api?
Yes. Audio tags in square brackets (for example: [sad], [angry], [whispers], [laughs]) are supported in the API when you use the Eleven v3 (Alpha) model. Include the tags inline in your text input and set model_id to "eleven_v3" with the Create speech or Stream speech endpoints; they’re also supported in the Text to Dialogue API. (help.elevenlabs.io)
For effective usage and a non‑exhaustive list of supported tags, see the Eleven v3 prompting guide. (elevenlabs.io)
Note: v3 is in public alpha and subject to the Beta Services Addendum restrictions called out in the help center article. (help.elevenlabs.io)
Sources:
How do audio tags work with Eleven v3 (Alpha)? – ElevenLabs
Prompting Eleven v3 (alpha) | ElevenLabs Documentation
What is Eleven v3 (Alpha)? – ElevenLabs
so can I used tags with the javascript api and eleven_v3 alpha?