Speech Service Providers¶
Speech service providers enable voice capabilities in Open Chat Studio, including text-to-speech (TTS) and speech-to-text (STT). Configuring a speech provider allows chatbots to synthesize spoken audio from text and transcribe audio input to text.
Supported providers¶
- ElevenLabs — TTS, STT, and voice management including custom voice cloning
- intron.io — TTS with 90 African and international accent voices
- MiniMax — TTS using MiniMax T2A voices
ElevenLabs¶
ElevenLabs is an AI audio platform offering high-quality voice synthesis and transcription.
Capabilities¶
| Capability | Details |
|---|---|
| Text-to-Speech (TTS) | Converts text to MP3 audio using ElevenLabs voices |
| Speech-to-Text (STT) | Transcribes audio using ElevenLabs Scribe v2 |
| Voice sync | Syncs available voices from the ElevenLabs catalog to OCS |
Setup¶
- Obtain an API key from your ElevenLabs account.
- In Open Chat Studio, navigate to Team Settings → Speech Service Providers.
- Add a new provider, select ElevenLabs, and enter your API key.
- On creation, OCS will automatically sync available voices from your ElevenLabs catalog.
Voices¶
When you create or edit an ElevenLabs provider, OCS fetches and stores the voices available in your ElevenLabs account. You can trigger a manual re-sync from the provider's edit page to pick up newly added voices.
Custom voices: Custom voices must be created directly in ElevenLabs. OCS does not support creating custom voices. Once created in ElevenLabs, they will be synced to OCS the next time a sync runs (automatically on provider creation, or manually on request).
OCS does not support creating custom voices directly — manage your voice library in ElevenLabs and let OCS sync the results.
intron.io¶
intron.io is a text-to-speech platform specialising in African language voices. It provides synthetic voices across a wide range of African and international accents, making it a strong choice for teams serving African-language audiences.
Capabilities¶
| Capability | Details |
|---|---|
| Text-to-Speech (TTS) | Converts text to audio using intron.io voices |
| Voice catalogue | 90 pre-seeded voices covering 45 accents × male and female variants |
Setup¶
- Obtain an API key from your intron.io account.
- In Open Chat Studio, navigate to Team Settings → Speech Service Providers.
- Add a new provider, select intron.io, and enter your API key.
- On creation, OCS will automatically seed all available voices for the provider.
Voices¶
When you create an intron.io provider, OCS automatically creates 90 voices — one male and one female variant for each of 45 accents. Example accents include Afrikaans, Hausa, Igbo, Kinyarwanda, Luganda, Swahili, Xhosa, Yoruba, and Zulu, alongside a further 36 African and international accents.
Voice seeding is idempotent: saving or re-saving the provider will not create duplicate voices.
MiniMax¶
MiniMax is an AI platform whose T2A (text-to-audio) API synthesizes spoken audio from text.
Capabilities¶
| Capability | Details |
|---|---|
| Text-to-Speech (TTS) | Converts text to audio using the MiniMax T2A v2 API |
| Voice catalogue | Built-in MiniMax voices seeded automatically on provider creation |
MiniMax provides text-to-speech only — it does not offer speech-to-text.
Setup¶
- Obtain an API key and Group ID from the MiniMax platform.
- In Open Chat Studio, navigate to Team Settings → Speech Service Providers.
- Add a new provider, select MiniMax, and enter your API key and Group ID. You can optionally set the T2A model; a sensible default is filled in for you.
- On creation, OCS automatically seeds the available MiniMax voices.
Voices¶
When you create a MiniMax provider, OCS automatically seeds its built-in MiniMax voices, so they are ready to select in your chatbots. Voice seeding is idempotent: saving or re-saving the provider will not create duplicate voices.