Neural Text-to-Speech

Neural TTS built in Europe.

First-chunk audio in under 150 ms. Natural voices in French and 40+ languages. Streaming synthesis for real-time products. EU-hosted, no US CLOUD Act exposure.

Get an API key Read the docs

EU production-gradeFirst chunk <150 msNeural voicesEU-hostedVoice cloning available

Capabilities

Natural voices, real latency, data in Europe.

First chunk under 150 ms

Streaming TTS delivers the first audio chunk in under 150 ms, enabling natural conversational pacing for real-time products and contact center flows.

Neural voices in 40+ languages

Natural-sounding speech synthesis with prosody tuned for European languages. French, German, Spanish, Italian, Portuguese, Dutch, and more.

Voice cloning available

Create a custom neural voice from as little as 60 seconds of audio. Available on enterprise contracts. Voice preservation on roadmap for late 2026.

EU-hosted, no transatlantic audio

Synthesis runs entirely inside European data centers (OVHcloud, Scaleway). No audio leaves the EU. GDPR-native by design.

Integration

One endpoint, streaming or batch.

curl https://api.voxist.com/v1/synthesize \
  -H "Authorization: Bearer $VOXIST_API_KEY" \
  -F text="Votre colis arrive demain." \
  -F voice=fr_neural_1 \
  -F format=wav

The /v1/synthesize endpoint supports streaming and batch output, with multiple output formats. Available as European SaaS, private cloud, or on-premise.

Explore the Voxist API platform