#6767th of 70 in Agent frameworks and SDKsFor science, health and research, research and papers
ai-scientist
Pipeline where LLMs propose ideas, run experiments and write LaTeX papers
- Stars
- 14.8k
- License
- custom license
- Last commit
- Dec 2025
- Language
- Jupyter Notebook
Missed the latest GitHub check; these facts may be older.
Overview
Runs an end-to-end loop in which an LLM generates research ideas, checks novelty via Semantic Scholar or OpenAlex, edits and runs experiment code, and writes a LaTeX paper. It also includes an LLM-based paper reviewer. Three templates ship (NanoGPT, 2D diffusion, grokking), plus community-contributed ones.
Who it is for: ML researchers experimenting with automated research agents
Strengths
- Covers idea generation, experiments, paper write-up and automated review in one pipeline
- Supports OpenAI, Anthropic (including Bedrock and Vertex), DeepSeek, OpenRouter and Gemini
- README cites typical cost under $15 per paper with Claude Sonnet 3.5
- Custom templates follow a documented file layout (experiment.py, plot.py, prompt.json)
Weaknesses
- Executes LLM-written code; README says to containerize and restrict web access yourself
- Needs Linux, NVIDIA GPU with CUDA and a full texlive install; CPU-only is impractical
- Ideas limited to those expressible in code; only three templates are maintained
- Custom license requires disclosing AI use in resulting manuscripts
What it needs
- GPU required
- Needs PyTorch, CUDA, texlive-full, conda, Semantic Scholar API (optional), OpenAlex (optional)
- Models: OpenAI GPT-4o, OpenAI o1, Claude Sonnet 3.5, DeepSeek, Llama 3.1 via OpenRouter
Also in Agent frameworks and SDKs
See all 70| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 94 out of 100 |
| 2 | langchainPython framework for building LLM agents and apps from interoperable components | 89 out of 100 |
| 3 | browser-usePython library that lets an LLM agent control a web browser | 89 out of 100 |
| #4 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 89 out of 100 |
| #5 | mastraTypeScript framework for building AI agents and workflows | 89 out of 100 |