ElevenLabs voice models on BazaarLink: 9 text-to-speech and 2 speech-to-text models, one API key
BazaarLink now offers ElevenLabs text-to-speech (Eleven v4, v3, Flash, Turbo, Multilingual) and speech-to-text (Scribe v2) through OpenAI-compatible /v1/audio endpoints. From $40 per million characters, with a listen-before-you-buy section on every model page.
The BazaarLink catalog now includes the ElevenLabs voice family: 9 text-to-speech (TTS) models and 2 speech-to-text (STT) models. They all use OpenAI-compatible endpoints, one BazaarLink API key and one invoice, priced in USD.
The models
Text to speech
| Model | Good for | Price (per million characters) |
|---|---|---|
elevenlabs/eleven-v4 | Newest flagship, most expressive | $80 |
elevenlabs/eleven-v3 | Rich emotion and tone | $80 |
elevenlabs/eleven-multilingual-v2 | Multilingual, stable | $80 |
elevenlabs/eleven-v4-turbo | Low-latency v4 | $40 |
elevenlabs/eleven-turbo-v2.5 | Low latency, multilingual | $40 |
elevenlabs/eleven-turbo-v2 | Low latency | $40 |
elevenlabs/eleven-flash-v2.5 | Fastest response, real-time conversation | $40 |
elevenlabs/eleven-flash-v2 | Fastest response | $40 |
elevenlabs/eleven-v3-conversational | Tuned for conversation | $40 |
Speech to text
| Model | Notes | Price |
|---|---|---|
elevenlabs/scribe-v2 | High-accuracy multilingual transcription | $0.0000611 per second (about $0.22 per hour) |
elevenlabs/scribe-v2-medical | Tuned for medical vocabulary | $0.0000611 per second (about $0.22 per hour) |
The price on each model page is the one that counts.
How to use them
Text to speech: /v1/audio/speech
curl https://api.bazaarlink.ai/v1/audio/speech \
-H "Authorization: Bearer $BAZAARLINK_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "elevenlabs/eleven-v4",
"voice": "george",
"input": "Welcome to BazaarLink!",
"response_format": "mp3"
}' \
--output speech.mp3
With the OpenAI Python SDK:
from openai import OpenAI
client = OpenAI(base_url="https://api.bazaarlink.ai/v1", api_key="sk-bl-YOUR_API_KEY")
with client.audio.speech.with_streaming_response.create(
model="elevenlabs/eleven-v4",
voice="george",
input="Welcome to BazaarLink!",
response_format="mp3",
) as response:
response.stream_to_file("speech.mp3")
Choosing a voice
Set voice to a voice name. Not sure which one fits? Open a model page and play the built-in voices in the listen section — samples in Chinese and English — then use that name in your request.
Speech to text
Transcription uses the OpenAI-compatible /v1/audio/transcriptions:
curl https://api.bazaarlink.ai/v1/audio/transcriptions \
-H "Authorization: Bearer $BAZAARLINK_API_KEY" \
-F "model=elevenlabs/scribe-v2" \
-F "file=@meeting.mp3"
Pricing in practice
- Text to speech: billed by the characters you send. 1,000 characters costs about $0.08 on Eleven v4 and about $0.04 on the Flash and Turbo models.
- Speech to text: billed by audio length; one hour of audio is about $0.22.
- You pay for what you use.
Listen first
Every text-to-speech model page has a listen section: play the built-in voices in Chinese and English without writing any code, then decide which model and voice to use. Signed-in accounts can also type their own text on the model page and generate audio on the spot.
→ Browse the models: https://bazaarlink.ai/models?q=elevenlabs
FAQ
How is it priced? See above: TTS by character, STT by second.
I already use the OpenAI SDK. Do I have to change much? No — change base_url and the API key.
TWD billing · Taiwan invoices · leading AI models · OpenAI-compatible API