A Claude Code-based job application framework that evaluates postings, tailors LaTeX CVs, writes cover letters and preps interviews, running locally in your fork.
55K17.7KPythonMITUpdated 2 days ago
aiming-lab/AutoResearchClaw
An open-source pipeline that turns a single research idea into a full academic paper, with literature search, sandbox experiments, citation checks and LaTeX output.
- Stars
- 14,719
- Forks
- 1,709
- Open issues
- 20
- Last commit
- 2 months ago
What is it?
AutoResearchClaw runs a 23-stage pipeline that takes a research topic and produces a draft paper, a conference-style LaTeX file, a BibTeX reference list, experiment code with metrics, charts, and multi-agent peer reviews. Literature is gathered from OpenAlex, Semantic Scholar and arXiv, and references are checked through a four-layer verification step meant to remove fabricated citations.
Experiments run in a sandbox or Docker environment, with hardware detection for NVIDIA CUDA, Apple MPS or CPU. Failed runs are diagnosed and repaired, and a decision stage can refine or pivot the research direction. Lessons from each run are stored and can feed later runs through the optional MetaClaw bridge.
The project can run fully autonomously from the CLI or in a Co-Pilot mode with human approval at chosen stages. It can also be driven by OpenClaw or any ACP-compatible coding agent such as Claude Code or Codex CLI.
Key features
- 01Collects real papers from OpenAlex, Semantic Scholar and arXiv
- 02Verifies citations with arXiv, CrossRef, DataCite and LLM checks
- 03Generates and runs experiment code in a sandbox or Docker with GPU/MPS/CPU detection
- 04Outputs LaTeX templates for NeurIPS, ICML and ICLR plus BibTeX
- 05Offers autonomous mode and human-in-the-loop Co-Pilot modes with approval gates
Use cases
- Drafting a first-pass paper from a new research idea to test its feasibility
- Automating literature review and experiment scaffolding for ML or science projects
- Studying how autonomous research agents behave across pipeline stages
- Running research tasks from a chat interface through OpenClaw or an ACP agent
Quick start
- 1
Clone the repository
git clone https://github.com/aiming-lab/AutoResearchClaw.git - 2
Enter the directory, create a virtual environment and install
cd AutoResearchClaw && python3 -m venv .venv && source .venv/bin/activate && pip install -e . - 3
Run interactive setup and configuration
researchclaw setup && researchclaw init - 4
Set your API key and run the pipeline on a topic
export OPENAI_API_KEY="sk-..." && researchclaw run --config config.arc.yaml --topic "Your research idea" --auto-approve
Things to know
- Needs an LLM API key (for example OpenAI) and typically Docker or LaTeX for full experiment and paper output.
- Fully autonomous runs can produce papers of uneven quality, so human review is advised; Co-Pilot mode is available for guidance.
Topics
FAQ
- What is AutoResearchClaw?
- An open-source pipeline that turns a single research idea into a full academic paper, with literature search, sandbox experiments, citation checks and LaTeX output. AutoResearchClaw runs a 23-stage pipeline that takes a research topic and produces a draft paper, a conference-style LaTeX file, a BibTeX reference list, experiment code with metrics, charts, and multi-agent peer reviews. Literature is gathered from OpenAlex, Semantic Scholar and arXiv, and references are checked through a four-layer verification step meant to remove fabricated citations.
- Is AutoResearchClaw free to use? What license does it have?
- AutoResearchClaw is open source under the MIT license. Check the license terms before using it in commercial products.
- Is AutoResearchClaw still maintained?
- Updated within 6 months. The last commit to AutoResearchClaw was on Aug 19, 2026.
- What can I use AutoResearchClaw for?
- Drafting a first-pass paper from a new research idea to test its feasibility Automating literature review and experiment scaffolding for ML or science projects Studying how autonomous research agents behave across pipeline stages Running research tasks from a chat interface through OpenClaw or an ACP agent
- How do I install AutoResearchClaw?
- Clone the repository: git clone https://github.com/aiming-lab/AutoResearchClaw.git Enter the directory, create a virtual environment and install: cd AutoResearchClaw && python3 -m venv .venv && source .venv/bin/activate && pip install -e . Run interactive setup and configuration: researchclaw setup && researchclaw init Set your API key and run the pipeline on a topic: export OPENAI_API_KEY="sk-..." && researchclaw run --config config.arc.yaml --topic "Your research idea" --auto-approve
- Where can I get AutoResearchClaw?
- The source code and installation instructions are on GitHub: https://github.com/aiming-lab/AutoResearchClaw
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A zero-code multi-agent platform where you define agents, workflows and tasks via YAML and a web console to build software and other outputs.
34.6K4.3KPythonApache-2.0Updated 4 days ago
A reference stack from NVIDIA for running AI agents such as OpenClaw, Hermes and LangChain Deep Agents more safely inside OpenShell sandboxes.
22.8K3.1KTypeScriptApache-2.0Updated 3 hours ago
Unity ML-Agents Toolkit lets Unity games and simulations serve as environments for training AI agents with reinforcement and imitation learning.
AI & Machine Learning#reinforcement-learning#unity#deep-reinforcement-learning#imitation-learning#game-ai
19.8K4.5KC#OtherUpdated 7 days ago
A zero-code platform that generates production-style AI agents from natural-language descriptions, with built-in tools, memory and orchestration.
6K744PythonMITUpdated 3 hours ago
A CLI and workflow engine that runs coding agents like Claude Code, Codex and Copilot through configurable review, repair and test gates before work is accepted.
2K189RustMITUpdated yesterday