Wallie V2: Open-Source AI Streaming Assistant with Screen Reaction and Live Chat Integration
By
Yelkhan
A bagel-shaped object. The form is there, the soul isn't.
Summary
Wallie V2 is an open-source AI streaming assistant that runs locally on your machine. It reacts to screen events using perceptual hashing (pHash) to detect meaningful changes, reads live chat from Twitch/YouTube/Kick, animates a Live2D avatar with real lipsync, and generates spoken reactions with 2-4 second latency. Users can freely swap LLM and TTS providers (e.g., Groq + Llama-4 Scout for speed, Claude Sonnet for quality). It starts free with Groq + Piper and has zero cloud lock-in.
Key quotes
· 4 pulledLatency from screen event to spoken reaction: typically 2–4 seconds end to end, depending on the LLM provider.
Groq + Llama-4 Scout gets you the fastest loop (~1.5–2s). Claude Sonnet is slower on raw latency but produces better reactions.
Wallie is an open-source AI streamer that actually feels alive. It reacts to your screen, reads live chat on Twitch/YouTube/Kick, animates a Live2D avatar with real lipsync, and never repeats itself — all running locally on your machine.
Start free with Groq + Piper. Zero cloud lock-in.
You might also wanna read
Kimi K2.5: Open-Source Multimodal AI Model with Visual Agentic Intelligence and Agent Swarm Capabilities
Kimi K2.5 is introduced as the most powerful open-source model to date, building on Kimi K2 with continued pretraining on approximately 15 t
Linum v2 Text-to-Video AI Model Released in 360p and 720p Versions
Linum-AI has released Linum v2, a text-to-video AI model available in two versions (360p and 720p) that generates 2-5 second videos. The mod
LemonSlice: Real-Time Video Platform for AI Voice Agents
LemonSlice is a real-time video platform for AI voice agents that enables them to see and interact with visual environments. The article dis
news.ycombinator.com·4mo ago
Building Ultra-Low-Latency Voice Agents with NVIDIA Open Models
This technical guide demonstrates how to build ultra-low-latency voice agents using NVIDIA's open models, including the newly launched Nemot
daily.co·4mo ago
Kuri: Zig-Native Browser Automation Tool for AI Agents with Token-Efficient CDP Snapshots
Kuri is a browser automation and web crawling tool specifically designed for AI agents, written in Zig programming language. It offers a lig
github.com·1mo ago
OpenAI's Sora 2 Video Generation Model and New Social App
Cal Newport analyzes OpenAI's new video generation model Sora 2, which creates realistic videos from text prompts, and discusses the accompa
calnewport.com·7mo ago