# bowang-lab/MedRAX

MedRAX: Medical Reasoning Agent for Chest X-ray - ICML 2025

Repository: https://github.com/bowang-lab/MedRAX
Canonical: https://ross.abutalabs.com/products/medrax
Homepage: https://bowang-lab.github.io/MedRAX/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent, agentic-ai, agentic-framework, chest-xrays, healthcare, langchain, medical-imaging, chest-xray, medical, reasoning-agent
Last push: 2025-10-31T00:31:34+00:00

## Health v2 (maintenance only)
Score: 42/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 49, release rhythm 35, longevity 40
- inputs: {"age_days": 573, "days_push": 307, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1218, forks 204 (observed 2026-08-28T04:04:01.524025+00:00)

## What it is
MedRAX is a medical reasoning agent framework that integrates chest X-ray analysis tools (segmentation, grounding, report generation, disease classification) with multimodal LLMs like GPT-4o into a unified orchestration layer built on LangChain and LangGraph. It ships with ChestAgentBench, a 2,500-query benchmark, and a Gradio-based interface for interactive CXR interpretation.

## Use cases
- interpret chest x-rays with an ai agent
- generate radiology reports from cxr images
- segment anatomical structures in chest x-rays
- classify lung pathologies in x-ray images
- benchmark medical ai agents on complex queries
- orchestrate medical imaging models with an llm
- localize findings in chest x-rays

## When to choose
- you need a unified agent to orchestrate multiple chest X-ray analysis models without training
- you want a research baseline or benchmark for medical agentic AI on CXR tasks
- you need DICOM processing plus segmentation, grounding, and reporting in one pipeline

## When to avoid
- you need a clinically certified diagnostic product for patient care
- your use case involves imaging modalities other than chest X-rays
- you want a lightweight model without GPT-4o API dependency or GPU-heavy submodels

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, machine-learning, image-processing, llm-inference, rag
- domain: healthcare, artificial-intelligence, computer-vision
- platform: python, self-hosted, cross-platform
- tags: chest-xray, medical-imaging, langchain, langgraph, multimodal, chestagentbench, report-generation, segmentation, gradio, research, ai-agents, docker

## Member repositories
- bowang-lab/MedRAX (main) score 42

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:01.524025+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:17:02.523647+00:00, confidence not recorded.
  - readme: https://github.com/bowang-lab/MedRAX (fetched 2026-08-28T04:04:01.524025+00:00, sha a11a67a331ca)
  - homepage: https://bowang-lab.github.io/MedRAX/ (fetched 2026-08-29T12:24:53.984836+00:00, sha 8382600dbaa5)
- Data as of 2026-08-30T08:39:29.467469+00:00.