Ultravox achieves open source SOTA in real-time voice AI. I've played with the demo and am VERY impressed. This is the closest we've come to OpenAI's Advanced Voice Mode yet (which still has some advantages). Even works really well in other languages (I tested it in German, too).

Talk to Ultravox (audio-native LLM) from Fixie.ai 🦊 on Hugging Face spaces! Powered by Gradio, webrtc, and transformers ⚡️

🌟 Roadmap to Human-Level Voice AI: The 5 Stages by Ultravox Fixie.ai 🦊 🔑 Key Insights: 💡 Assistant (Level 1): Basic task execution and information retrieval, like Siri or Alexa, with limited ability to understand context or sustain multi-turn conversations. 🗣️

Nice to see OpenAI continuing to follow in our footsteps! :) I'll just point out that: * We're still cheaper * We're still open * We have more voices * We're just as fast (if not faster!)

import { UltravoxSession } from 'ultravox-client'; const session = new UltravoxSession(); session.joinCall(); Yours is good too, tho

Clearer conversations, smarter communication! 💬✨ Krisp is partnering with Fixie.ai to enhance Ultravox Realtime with AI Noise Cancellation, ensuring seamless Voice AI conversations anywhere and noise-free. 🚀 Learn more: 👉 krisp.ai/blog/krisp-ext…

Watch our CEO Zach Koch talk with Davit @Krisp from Krisp about the future of Realtime Voice AI and what we're doing with Ultravox to make it possible

Chainlit Spree Commerce Lon Ligrin NautilusTrader vik 🚀Project Number 9 - Ultravox🔥 'The Next Generation of Multimodal AI for Real-Time Voice Interaction' by @fixieai #opensource #github #trending #AI #coding #search #softwaredevelopment #technology #innovation #programming #AItools #Ultravox

Building real-time WebRTC and Websocket applications is very difficult to get right in Python. Until now - Introducing FastRTC, the realtime communication library for Python ⚡️

To get started pip install fastrtc Join and Follow our org: huggingface.co/fastrtc Check out the cookbook fastrtc.org/cookbook/ for conversational apps powered by Google Gemini ✨ , SambaNova Systems, ElevenLabs, Groq Inc , Fixie.ai 🦊, kyutai , OpenAI File requests and

so, the OpenAI realtime API is actually probably the worst realtime audio option on the market. Gemini's multimodal beta blows it out of the water in terms of voice quality, intelligence, and instruction following, and Fixie.ai 🦊 / Ultravox kicks ass too. gpt-4o-realtime-preview

It's been great working with the 11x team to help bring Julian to life (built on top of Ultravox!). Check him out:

⚡ Voice AI that responds in 600ms? Meet Ultravox — a multimodal LLM from Fixie.ai 🦊 that skips the STT step and processes audio directly into an LLM. Built on cerebriumai infra, deployed with Pipecat from Daily

kwindla I’m using Ultravox from Fixie.ai 🦊 to build AI agents for call centers and the healthcare industry, and for me, it’s the best. They offer very advanced tool calls and a design pattern called “stages,” similar to agent orchestration in traditional LLMs. Plus, you can manage state

Voice AI, the enterprise way. Hexaware partners with Fixie.ai 🦊 to deliver faster, sharper customer conversations using Ultravox Realtime. Deployed where accuracy, scale, and uptime are non-negotiable. Read more: bit.ly/4kE7y8h #VoiceAI #EnterpriseAI #UltravoxRealtime