aiming-lab/AutoResearchClaw

Fully autonomous & self-evolving research from idea to paper. Chat an Idea. Get a Paper. 🦞

What it solves

AutoResearchClaw is an autonomous research pipeline designed to transform a single research idea into a conference-ready academic paper. It eliminates the manual drudgery of literature review, experiment design, code execution, and LaTeX drafting, while providing a system to prevent common AI issues like hallucinated citations and fabricated results.

How it works

The system operates through a 23-stage pipeline divided into eight phases, ranging from initial topic scoping and literature discovery (using real APIs from OpenAlex, Semantic Scholar, and arXiv) to experiment execution in a hardware-aware sandbox and final LaTeX export. It can run in two primary modes:

  • Fully Autonomous: The pipeline executes all stages independently, including self-healing experiments and autonomous pivoting if hypotheses fail.
  • Co-Pilot Mode: A human-in-the-loop system with six intervention modes (e.g., step-by-step, co-pilot) allowing users to guide hypothesis creation, review experiment designs, and co-write the paper.

The project integrates with various AI coding agents via the Agent Client Protocol (ACP) and can be bridged to messaging platforms like Discord and Telegram via OpenClaw.

Who it’s for

It is designed for researchers and academics who want to accelerate the research cycle, from those seeking a fully automated draft to those who want an AI partner to help navigate the experimental and writing phases.

Highlights

  • Multi-Domain Execution: Includes specialist agents for high-energy physics, biology, and statistics, alongside a generic Docker executor.
  • Anti-Fabrication: Features a Sentinel Watchdog and a 4-layer citation integrity system to verify claims against real literature.
  • Self-Evolving: Uses MetaClaw to extract lessons from failed runs to improve future performance.
  • Conference-Ready Output: Generates full academic papers in LaTeX targeting templates for NeurIPS, ICML, and ICLR.
  • Flexible Integration: Compatible with multiple LLM backends (Claude Code, Gemini CLI, etc.) and messaging platforms.

Related

  • Project
    federicodeponte/opendraft

    An open-source AI research engine that generates source-grounded academic drafts with citations verified against multiple scholarly databases to prevent hallucinations.

  • Project
    wentorai/Research-Claw

    A local AI agent system that automates academic research workflows, including literature management, paper drafting, and autonomous experimental reproduction.

  • Project
    OpenLAIR/dr-claw

    Dr. Claw is an open‑source, model‑agnostic AI research assistant that provides a unified UI/CLI for the whole research workflow—literature survey, idea generation, experiment code, analysis, paper drafting, and slide creation—driven by Claude Code, Gemini, Codex or any OpenRouter model.

  • Project
    Ar9av/PaperOrchestra

    A pluggable skill pack for coding agents that implements a multi-agent pipeline to transform unstructured research logs and ideas into professional LaTeX research papers.

  • Project
    dramaclaw/dramaclaw

    An industrialized AI production pipeline that transforms manuscripts into finished short dramas by automating script generation, storyboarding, voice-over, and video composition.