Scott Stagg
Principal Investigator & Director
Florida State University · SECM4
Cryo-EM methodology, high-throughput structural biology, COPII vesicles.
About
Built by scientists, for science.
Magellon is open-source cryo-EM software built by structural biologists and software engineers who use cryo-EM in their own work. The goal is to make the path from a fresh micrograph to a usable result short, reproducible, and pleasant — without locking anyone into a particular cluster, vendor, or workflow.
open source · apache 2.0 cloud native plugin platform
Vision
Decoupled geographically. Integrated operationally.
We want cryo-EM data collection and analysis decoupled geographically while remaining integrated operationally. Magellon lets researchers reach world-class microscopes and computational resources without owning the hardware, while keeping a single interface that hides the distributed-computing details.
Plugins are first-class. Drop in a new algorithm, declare its inputs and outputs as Pydantic schemas, and the UI renders the form, runs the job, and shows the result alongside everything else — no per-plugin frontend work required.
The team
Researchers and engineers, sharing a stack.
The PIs bring expertise in membrane trafficking, AAA+ proteases, microtubule transport, and cryo-EM methodology. The development team focuses on the platform — making the science accessible through software that scales from a laptop to a cluster.
Gabe Lander
Principal Investigator
Scripps Research Institute
AAA+ proteases, protein degradation, advancing cryo-EM accessibility.
Michael Cianfrocco
Principal Investigator
University of Michigan · LSI
Microtubule transport, cryo-EM automation, cloud computing for structural biology.
Behdad Khoshbin
Lead Developer
Project architecture
System design, plugin platform, frontend & backend architecture.
Puneeth Reddy
Developer
Florida State University
Image processing, frontend components, scientific computing.
The stack
Containerized end-to-end.
Magellon ships as a set of Docker images and one start script. The pieces that matter to most users:
Frontend
-
React 19
TypeScript, Material UI, Zustand state
-
Socket.IO
Live job progress streaming
Backend
-
FastAPI
Python, SQLAlchemy ORM
-
magellon-sdk
Pydantic schemas → auto-generated UI forms
Data plane
-
MySQL 8
Sessions, images, jobs, results
-
RabbitMQ
Plugin task queues
-
NATS
Live event streaming
-
GPFS root
MRC / TIFF / JPEG image files on disk
Delivery
-
Docker
Single start.sh brings the full stack up
-
Apache 2.0
Open source, free for any use
Get started
Want to dig deeper?
The full architecture lives in the docs — installation, configuration, plugin authoring, and the API reference.