✨ Today we’re releasing JoyAI-VL-Interaction — our fully open-source real-time vision-language interaction model and system. 🎉 Last month, TML introduced the idea of an “interaction model.” We were exploring a similar direction concurrently, and now we’re open-sourcing the full stack 🧠 Unlike turn-based models, it watches the world with you and decides every second whether to speak, stay silent, or delegate harder tasks to a background agent It can: 🔥 Alert instantly when fire breaks out or someone falls 🎯 Count, translate, and commentate in real time 📱 Guide you step by step through app interactions 🤖 Delegate complex digital tasks via OpenClaw or Claude Code while continuing to watch A few things surprised even us: 🤯 With zero app-interface videos in training, it still learned to guide phone-screen shopping flows ⚡ Its timing is what really stands out: an 8B model can speak up at the right moment, while much larger models often rely on external polling and arrive a beat late 🌱 With just 4M samples, it already begins to show an “EQ” for when to respond We think interaction models may be the next paradigm: models that are truly present in real-time and able to bridge the physical and digital worlds. 🚀 Play and build with us: 🏆 Blog: joyai-vl-video-future-academ… 🤗 HuggingFace: huggingface.co/papers/2606.1… 💻 GitHub: github.com/jd-opensource/Joy…

Thanks for tweeting our work🍻

Thanks so much for tweeting our work!

Many thanks for your tweeting 👏

🚀 We’re excited to announce the release of InternLM-XComposer2.5-OmniLive (IXC2.5-OL), a comprehensive multimodal system designed for long-term streaming video and audio interactions. This fully open-sourced project delivers functionality similar to Gemini 2.0 Live Streaming and OpenAI Her, with standout features including: 🎥 Chat with Streaming Video & Audio 💾 Long-Term Memory for recalling past video experiences 🏆 Competitive Performance across various video and audio perception benchmarks 📄 Paper: huggingface.co/papers/2412.0… 💻 Code: github.com/InternLM/InternLM… 📦 Models: huggingface.co/internlm/inte… ✨ Immerse yourself in multimodal interaction and create your own app today! #Gemini2 #OpenAI #ChatGPTAdvancedVoice