Marlin-2B: a tiny VLM to extract structured information from videos
Beats Qwen2.5-VL-7B on temporal grounding while running on a single consumer GPU.
True AI-Native Browser — a VLM reads the HTML and hallucinates the page.
Ditches the rendering engine entirely to let a VLM hallucinate the pixels from HTML.
AI researchers, creative coders, and browser engine enthusiasts
Arc Browser · Comet Browser · Standard Web Browsers
Beats Qwen2.5-VL-7B on temporal grounding while running on a single consumer GPU.
Daily arXiv scraping with Claude classification beats manual curation.
3-line real-time VLM API, but competing products handle camera inference already.
Uses Google's CLD3 neural net to match Chrome's native API behavior exactly.
355 LOC Chromium hack for Figma-like HTML performance, but security concerns unaddressed.
Proxies gopls for seamless IDE navigation without pointing to generated files.