loopless
Built at WeaveHacks 3: Self-Improving Agents Hackathon with Weights & Biases · Jan 31, 2026 · San Francisco, CA

LoopLess is a self-improving browser automation agent. It uses BrowserBase for cloud browser sessions with live streaming and permanent recordings. Gemini powers the LLM planning and serves as an LLM-as-a-Judge for evaluating task completion. W&B Weave provides full observability with tracing, custom scorers, and the built-in Evaluation framework to analyze failures and feed learned rules back into prompts. Redis caches successful action sequences as "macros"—cold runs learn, warm runs replay with zero LLM calls for 30-50% faster execution. The self-improvement loop automatically detects patterns (loops, missed actions) and improves future runs. Includes 18 AGI Inc benchmark tasks with real-time UI monitoring.