snap-stanford

Biomni: Run Biomedical Research Tasks with an AI Agent

Biomni is a Python biomedical AI agent that combines large language model reasoning, retrieval-augmented planning, and code execution to help researchers carry out tasks across biomedical fields. It also offers a web interface and supports external tools through MCP.

QuietPythonApache-2.0AI AgentsAIBiomedicine
Stars
4k
Forks
728
Stars, last 30 days
+103
Last commit
2 days ago
Contributors
40

Our take

Quietno commits in the last 90 days

Health score

36/100

How it is scored
Activity
31
Community
72
Issues
21
Pull requests
20

+103 stars in the last 30 days

Biomni is a broad biomedical research agent with useful integrations, but its current quiet development status and unrestricted execution of generated code call for careful, isolated evaluation.

Good fit if

  • You are a biomedical researcher or developer exploring LLM-assisted, multi-step research workflows.
  • Your team can set up the Python environment, configure an LLM provider, and accommodate the data lake when needed.
  • You want to extend an Apache-2.0 project with biomedical tools, data, or know-how contributions.

Look elsewhere if

  • You need a narrowly scoped analysis tool with a smaller setup footprint than Biomni's environment and optional large data lake.
  • You cannot run LLM-generated code with access to files, networks, and system commands, even in an isolated environment.
  • You require evidence of frequent recent maintenance or issue and pull-request handling before adopting a research tool.
All health signals
Last commit2026-10-19 (2 days ago)
Commits, last 90 days0
Releases, last 12 months1 (latest v0.0.8, 2025-10-27)
Contributors40 (top contributor: 35% of commits)
Issues closed, last 90 days0
Pull requests merged, last 90 days0
Project age19 months

Checked on 2026-10-22 with the GitHub API.

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

4,092

Forks

745

Change since Nov 17, 2026: +3 stars·0 forks

Overview

Biomni is a general-purpose biomedical AI agent built on large language model reasoning, retrieval-augmented planning, and code-based execution. It aims to help scientists carry out research tasks across biomedical fields and generate testable hypotheses, rather than serving as a single-purpose analysis tool.

The project is available as a Python package and includes a web interface. It can draw on a biomedical data lake and a curated know-how library, and can connect to external tools through MCP. The repository also includes Biomni-R0, a biology reasoning model, and Biomni-Eval1, a benchmark for biological reasoning.

Key Features

  • Execute biomedical tasks from natural-language instructions through the A1 agent.
  • Use retrieval and code execution to plan and carry out multi-step research workflows.
  • Retrieve protocols, best practices, and troubleshooting guidance from the know-how library.
  • Connect MCP servers to add external tools.
  • Launch an interactive Gradio interface or use the hosted web interface.
  • Save agent conversation histories as PDF reports.
  • Evaluate biological reasoning with the Biomni-Eval1 benchmark.

Use Cases

  • Biomedical researchers can ask the agent to help design a CRISPR screen or develop testable hypotheses.
  • Computational biologists can use it to support single-cell RNA-seq annotation workflows.
  • Drug discovery teams can request predictions such as compound ADMET properties.
  • Developers can connect external biomedical services through MCP when the built-in capabilities do not cover a task.

What you need

Detected in the repository

  • Python >=3.11 (from pyproject.toml)

License in plain words

Apache-2.0permissive

  • Commercial use: yes
  • Modify and redistribute: yes
  • You must keep: the license, the NOTICE file and a note of your changes
  • Share your changes: no
  • Includes an explicit patent grant from the contributors:

A summary, not legal advice: the LICENSE file is what applies.

Getting Started

Set up the environment following the environment guide, then install the package:

conda activate biomni_e1
pip install biomni --upgrade

Configure an LLM API key and see the README for agent examples, data lake options, and configuration details.

Alternatives

  • asta-paper-finder: PaperFinder focuses on finding and ranking research papers, while Biomni supports a broader range of biomedical research tasks through tools and code execution.
ProjectLanguageLicenseStarsStatus
BiomniPythonApache-2.04kQuiet
asta-paper-finderPythonApache-2.0251Maintained

Considerations

  • The project is marked quiet, with no commits in the measured 90-day period and no issues closed or pull requests merged during that period. It has 40 contributors, and the measured contributor share does not indicate dependence on one person.
  • The repository requires Python 3.11 or later. The README describes the software environment as large, and the agent downloads about 11 GB of data lake files by default unless that behavior is disabled.
  • Biomni executes LLM-generated code with full system privileges. The authors advise using isolated or sandboxed environments and caution against exposing sensitive data or credentials.
  • Biomni is Apache-2.0 licensed, but some integrated tools, databases, and software may have more restrictive licenses. Review those components before commercial use.
  • The README notes package dependency conflicts that may require manual installation or code changes, and says the included release was frozen as of April 15, 2025, differing from the current web platform.

Found this useful?

Share it with someone who would like Biomni.

Comparisons