The AI that catches your Eureka moments.
Crawls arXiv · Generates theorems · Proves lemmas · Writes LaTeX papers · Runs experiments
All from your chat or terminal.
English | 中文
$ eurekaclaw prove "Find recent papers on sparse attention + prove efficiency bound"
🦞 Crawling arXiv cs.LG (2024–2025)...
📄 Found 23 relevant papers. Summarizing...
💡 Hypothesis generated: O(n log n) via topological filtration
✨ Theorem 3.1 drafted. LaTeX ready. Proof complete.
🦞 Eureka! Paper draft saved to ./results/
EurekaClaw is a multi-agent AI research assistant that goes from a question to a publishable result — autonomously. It crawls the literature, generates and stress-tests hypotheses, runs experiments, and writes up findings, all from your terminal or browser UI.
Open Source · Local-First · Privacy by Design · Apache 2.0 License
| Feature | Description | |
|---|---|---|
| 🔍 | Literature Crawler | Fetch, summarize, and cross-reference papers from arXiv and Semantic Scholar |
| 💡 | Idea Generator | Brainstorm novel hypotheses by synthesizing patterns across thousands of papers |
| 🔢 | Theorem Prover | Generate, verify, and formalize proofs via a 7-stage bottom-up pipeline |
| 📄 | Paper Writer | Draft camera-ready LaTeX papers with theorem environments and citations |
| 🖥️ | Runs Locally | Compatible with Every Major Model API — Privacy by Design |
| 🧠 | Continual Learning | Distills proof strategies into skills after every session, improving over time |
| 🧪 | Experiment Runner (under development) | Numerically validates theoretical bounds; flags low-confidence lemmas |
| 🌐 | Browser UI | React + TypeScript interface — live agent track, proof sketch, pause/resume, skills manager |
(See Installation for detailed instruction)
macOS / Linux
curl -fsSL https://eurekaclaw.ai/install.sh | bashWindows
powershell -c "irm https://eurekaclaw.ai/install_win.ps1 | iex"The macOS/Linux installer clones the repo, creates a virtual environment, installs EurekaClaw, and adds the eurekaclaw command to your PATH. Run eurekaclaw onboard afterwards to configure your API key and settings.
Manual install (all platforms)
Requirements: Python ≥ 3.11, Node.js ≥ 20, Git
- Linux/MacOS
git clone https://github.com/EurekaClaw/EurekaClaw
cd EurekaClaw
make install # pip install -e "." + npm install (frontend)- Windows
git clone https://github.com/EurekaClaw/EurekaClaw
cd EurekaClaw
powershell -ExecutionPolicy Bypass -File install_win.ps1 # pip install -e "." + npm install (frontend)eurekaclaw onboard # interactive setup wizard (creates .env)
# — or — cp .env.example .env and add ANTHROPIC_API_KEY manually
eurekaclaw install-skills # install built-in proof skills (do once)
# Browser UI — build frontend and open in browser
make open
# CLI — prove a conjecture
eurekaclaw prove "The sample complexity of transformers is O(L·d·log(d)/ε²)" \
--domain "ML theory" --output ./results
# CLI — explore a domain
eurekaclaw explore "multi-armed bandit theory"
# CLI — start from arXiv papers
eurekaclaw from-papers 1706.03762 2005.14165 --domain "attention mechanisms"
# Browser UI - recommended
eurekaclaw ui --open-browserNo API key? Use a Claude Pro/Max subscription via OAuth.
| Command | Level | When to use |
|---|---|---|
eurekaclaw prove "<conjecture>" |
1 | You have a precise mathematical statement to prove |
eurekaclaw from-papers <ids> |
2 | You want to extend or find gaps in specific papers |
eurekaclaw explore "<domain>" |
3 | You have a broad research area but no conjecture yet |
See detailed documentation in https://eurekaclaw.github.io/ .
| 📖 User Guide | Installation, walkthrough, gate modes, tuning, troubleshooting |
| ⚙️ Configuration | All .env variables with defaults |
| 🏗️ Architecture | Pipeline stages, data flow, component design |
| 🤖 Agents | Each agent's role, inputs, outputs, and tool usage |
| 🔧 Tools | arXiv, Semantic Scholar, Lean4, WolframAlpha, code execution |
| 💻 CLI Reference | All commands and options |
| 🐍 Python API | EurekaSession, KnowledgeBus, data models |
| 🧠 Memory System | Episodic, persistent, and knowledge graph tiers |
| ✨ Skills | Skill registry, injection, distillation, writing custom skills |
| 🔌 Domain Plugins | Plugin architecture, MAB domain, adding new domains |
| 🌐 UI Design | React/TS architecture, component tree, run commands |
cp .env.example .env| Variable | Default | Description |
|---|---|---|
ANTHROPIC_API_KEY |
— | API key (or use OAuth, see User Guide) |
EUREKACLAW_MODEL |
claude-sonnet-4-6 |
Main reasoning model |
GATE_MODE |
auto |
none · auto · human |
THEORY_PIPELINE |
default |
default or memory_guided |
OUTPUT_FORMAT |
latex |
latex or markdown |
EXPERIMENT_MODE |
auto |
auto · true · false |
THEORY_MAX_ITERATIONS |
10 |
Max proof loop iterations |
Full reference → configuration.md
EurekaClaw includes a Scientist-Bench evaluator:
| Dimension | Weight |
|---|---|
| Formal correctness (Lean4 / LLM peer review) | 0.35 |
| Novelty (embedding distance from known results) | 0.25 |
| Experimental alignment | 0.15 |
| Proof depth (lemma count) | 0.15 |
| Citation coverage | 0.10 |
eurekaclaw eval-session <session_id># Unit tests (no API key needed)
pytest tests/unit/ -v
# Integration tests
ANTHROPIC_API_KEY=sk-... pytest tests/integration/ -v
# Frontend type-check
make typecheck
# Frontend development (hot-reload)
make devTo add a custom skill, drop a .md file into ~/.eurekaclaw/skills/ — see Skills and Continual Learning.
To add a new research domain, subclass DomainPlugin — see Domain Plugin System.
To add a new tool, subclass BaseTool and register it — see Research Tools.
EurekaClaw builds on ideas and inspiration from the broader AI-for-science community. We thank the authors of the following projects:
- MetaClaw — multi-agent research orchestration
- AutoResearchClaw — automated research orchestration
- EvoScientist — evolutionary hypothesis generation
- AI-Researcher — automated research pipeline
- Awesome AI for Science — curated resource list
- Dr. Claw — open research agent framework
- OpenClaw — open-source research claw
- ClawTeam — collaborative research agents
- ScienceClaw — science-focused research agent
If you use EurekaClaw in your research, please cite:
@misc{eurekaclaw2026,
title = {EurekaClaw: An AI Agent for Capturing Eureka Moments},
author = {Li, Xuheng and Di, Qiwei and Zhang, Chenggong and Ji, Kaixuan and Zhao, Qingyue and Liu, Yifeng and Zhang, Shiyuan and Gu, Quanquan},
year = {2026},
url = {https://github.com/EurekaClaw/EurekaClaw}
}Apache 2.0 License. See LICENSE for details.
Built for researchers who believe the next breakthrough is one Eureka moment away. 🦞