GitHub – aiming-lab/AutoResearchClaw: Fully autonomous research from idea to paper. Chat an Idea. Get a Paper. Fully Autonomous. 🦞

**Chat an idea. Get a paper.**

That’s the promise behind AutoResearchClaw, an open source project that aims to take you from a rough research thought to a full academic paper, without the usual marathon of tabs, drafts, rewrites, and late night debugging.

If you’ve ever tried to turn a spark of curiosity into something publishable, you know the process. First comes excitement. Then literature review overload. Then broken experiments. Then formatting LaTeX at 2 a.m. while questioning your life choices.

AutoResearchClaw tries to handle all of that for you.

You give it a topic. It searches real literature from arXiv and Semantic Scholar. It designs and runs experiments, automatically detecting whether you’re on GPU, MPS, or CPU. It performs statistical analysis. It even runs multi agent peer review before generating conference ready LaTeX targeting venues like NeurIPS, ICML, or ICLR.

And it doesn’t just stop when something fails. If experiments break, it self heals. If a hypothesis collapses, it pivots. If citations turn out to be fake, it removes them. That last part alone feels like a quiet revolution in a world worried about hallucinated references.

What I find especially interesting is the structured pipeline. There are approval gates where you can step in, or let it auto approve. There are decision loops that refine or pivot the direction of the research. It feels less like a simple script and more like a research lab condensed into software.

You can run it standalone, through CLI, or integrate it with OpenClaw for a more conversational workflow. One message, and the pipeline starts.

If you’re curious, you can explore the project here:
https://github.com/aiming-lab/AutoResearchClaw

We’re still at the beginning of autonomous research systems. They won’t replace human curiosity anytime soon. But they might become powerful collaborators, the kind that handle the heavy lifting so you can focus on asking better questions. And honestly, that shift alone could change how we do science.