Biomni  by snap-stanford

Biomedical AI agent for autonomous research tasks

created 4 months ago
1,905 stars

Top 23.4% on sourcepulse

GitHubView on GitHub
Project Summary

Biomni is a general-purpose AI agent designed to automate complex biomedical research tasks, from planning experiments to analyzing data. It targets researchers seeking to accelerate discovery and generate novel hypotheses by leveraging LLM reasoning, retrieval-augmented planning, and code execution.

How It Works

Biomni integrates LLM reasoning with retrieval-augmented planning and code execution to autonomously perform biomedical research tasks. This approach allows it to interpret natural language requests, break them down into actionable steps, retrieve relevant information, and execute code to achieve the desired outcome, enhancing research productivity.

Quick Start & Requirements

  • Install via pip install biomni --upgrade or from source: pip install git+https://github.com/snap-stanford/Biomni.git@main.
  • Requires setting API keys for Anthropic (Claude) and optionally OpenAI or AWS Bedrock in ~/.bashrc.
  • A setup script (setup.sh) is provided for environment setup, followed by conda activate biomni_e1.
  • Initial data download is approximately 11GB.
  • Official web interface available at biomni.stanford.edu.

Highlighted Details

  • Executes LLM-generated code with full system privileges; use in sandboxed environments is recommended.
  • Supports multiple LLMs including Claude and AWS Bedrock models.
  • Facilitates contributions of new tools, datasets, software, and benchmarks.
  • Actively developing Biomni-E2 for an expanded biomedical action space.

Maintenance & Community

Biomni is an open-science initiative with a call for community contributions. Details on contributing and upcoming features are available.

Licensing & Compatibility

Biomni itself is Apache 2.0-licensed. However, integrated tools, databases, or software may have more restrictive commercial licenses, requiring careful review for commercial use.

Limitations & Caveats

The current release (frozen April 15, 2025) executes code with full system privileges, posing a security risk for sensitive data or credentials. Some integrated components may have commercial use restrictions.

Health Check
Last commit

22 hours ago

Responsiveness

Inactive

Pull Requests (30d)
52
Issues (30d)
54
Star History
1,915 stars in the last 90 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems) and Omar Sanseviero Omar Sanseviero(DevRel at Google DeepMind).

scikit-llm by BeastByteAI

0.1%
3k
SDK for integrating LLMs into scikit-learn pipelines
created 2 years ago
updated 2 days ago
Feedback? Help us improve.