Deepgram Nova-3 + Aura-2 + Flux Multilingual

Voice & Multimodal

Deepgram Nova-3 provides speech-to-text in 45+ languages, with conversational STT and mid-call language switching across 10 languages, while Aura-2 delivers sub-200ms text-to-speech. Flux Multilingual supports multilingual processing for combined speech AI workflows.

🌐 Visit website ↗

🔗 Similar tools

Deepgram Voice Agent API

Single endpoint combining Nova-3 speech-to-text, LLM routing, Aura-2 text-to-speech, and F…
ElevenLabs Eleven v3 + ElevenAgents

ElevenLabs Eleven v3 delivers text-to-speech in 70+ languages with emotional Audio Tags, w…
Whisper v3 (local) ★ 108.5k

Whisper v3 is a local, open-source automatic speech recognition model trained on 680,000 h…
Jina-embeddings-v3 ★ 21.9k

Jina-embeddings-v3 is a multilingual text embedding model built on Jina-XLM-RoBERTa, suppo…

RAG & Memory
Pipecat ★ 15.2k

Pipecat is an open-source framework (BSD-2-Clause) for building voice and multimodal conve…
Kokoro ★ 8.7k

Kokoro is an open-weight text-to-speech model with 82 million parameters, Apache-licensed,…