Gradium is now available on Eden AI
Eden AI is an AI gateway: one API for 500+ models across LLMs, speech, vision and translation, with routing, fallbacks and a dedicated EU endpoint for European data residency. Gradium's Text-to-Speech, Speech-to-Text and speech translation are now available through that API.
A voice recording identifies the speaker and, in support, healthcare or banking calls, often carries personal data. Teams building voice products have to know where that audio is processed and which company processes it.
A European voice stack on Eden AI's EU endpoint
Eden AI customers asked for European voice options. Until now, Text-to-Speech on Eden AI came from US providers, which left teams that want voice data processed in Europe, by a European company, with no choice.
Gradium is built in Paris, and Eden AI calls to Gradium are pinned to the EU region, so teams get Gradium's voice quality and European data residency through the API key they already use.
Why Gradium runs as a passthrough on Eden AI
Eden AI maps most providers to a shared schema: one request shape, one response shape, and switching providers means changing one string. That mapping adds a few milliseconds per call. Batch transcription absorbs them, but less so in a real-time voice agent.
Gradium's TTS reaches first audio in 50ms, so a normalization layer would add latency in the same order of magnitude as the model itself.
Gradium therefore runs as a passthrough on Eden AI. You keep the single key, the single bill and the rest of the catalogue, and Text-to-Speech returns audio in the same request.
How to call Gradium TTS, STT and translation through Eden AI
import requests
response = requests.post(
"https://api.edenai.run/v3/universal-ai",
headers={"Authorization": "Bearer YOUR_EDENAI_API_KEY"},
json={
"model": "audio/tts/gradium",
"input": {"text": "Bonjour, comment puis-je vous aider ?"},
},
)Speech-to-Text and speech translation use the same shape on audio/speech_to_text_async/gradium and audio/speech_to_text_async/gradium/stt-translate. On Eden AI, these two run as async jobs, suited to recorded audio. For live streaming transcription, call the Gradium API directly.
Get started
With Gradium on Eden AI's EU endpoint, the audio is processed in Europe by a European company. To try it, grab an Eden AI key and set model to audio/tts/gradium.
- Gradium on Eden AI: edenai.co/providers/gradium
- Eden AI documentation: docs.edenai.co
- Gradium documentation: docs.gradium.ai
- Questions or volume pricing: talk to us
Frequently Asked Questions
Post to https://api.edenai.run/v3/universal-ai with your Eden AI key and set model to audio/tts/gradium for Text-to-Speech. Speech-to-Text and speech translation use the same shape on audio/speech_to_text_async/gradium and audio/speech_to_text_async/gradium/stt-translate.
In Europe. Eden AI calls to Gradium are pinned to the EU region, and Gradium is a European company built in Paris, so the audio is processed in Europe by a European company.
Eden AI maps most providers to a shared schema, which adds a few milliseconds per call. Gradium's TTS reaches first audio in 50ms, so a normalization layer would add latency in the same order of magnitude as the model itself. As a passthrough, you keep the single key, the single bill and the rest of the catalogue, and Text-to-Speech returns audio in the same request.
On Eden AI, Speech-to-Text and speech translation run as async jobs, suited to recorded audio. For live streaming transcription, call the Gradium API directly.