AutoResearchClaw
An autonomous multi-agent research pipeline that takes a research idea through experiment design, execution and paper writing, with an optional human-in-the-loop co-pilot mode.
AutoResearchClaw is a self-evolving pipeline for autonomous scientific research: given an idea, it collaborates across multiple agents to design experiments, run them and draft a paper, and has produced papers across domains including math, statistics, biology, computing, NLP, reinforcement learning, vision and robustness.
A domain-specialist execution stage routes experiments to specialist agents, such as a physics agent that goes from a Lagrangian through simulation tools, a biology agent using genome-scale metabolic modelling, and a statistics simulation-study agent, with a generic Docker executor for other domains, plus an accompanying benchmark of research topics with graded rubrics.
A human-in-the-loop system adds intervention modes ranging from fully autonomous to step-by-step review, along with tools for co-writing hypotheses and papers, confidence-driven pauses, and claim verification against hallucination.