Transform your data curation, analysis, and collaboration

Flywheel Enterprise streamlines imaging research with a medical imaging platform that automates time-consuming processes and scales for your needs. Our secure data management and AI platform gives you the power to create analysis-ready datasets for algorithm training and development.

Schedule A Demo

Accelerate productivity and improve collaboration

Ingest data from multiple sources and curate it to common standards.
Automate processing and machine learning pipelines.
Collaborate securely while maintaining compliance.

The Modern Medical Imaging Platform for Data and AI

Data Capture

Streamline capture of medical imaging and associated data

- Designed for multi-modality biomedical research data management
- Manages a wide array of data types including imaging files, clinical EMR data, instruments, and more
- Integrate data from external research platforms
- Option to add custom data types via Gears and viewers
- Rapid and automated data capture for both retrospective and prospective data utilizing connectors for imaging modalities, PACS, RIS, raw data, or file system
- Tools for high-performance bulk loading
- Simple drag-and-drop web-upload with no software installation required
- Solutions for HIPAA and 21 CFR Part 11 compliant medical research data management
- Configurable de-identification capabilities available through all capture tools
- Set unique whitelists, blacklists, and other rules (remove, replace, date processing, etc.) to ensure data is compliant and ready for research
- De-identification support prior to upload or between Flywheel projects to meet specific privacy needs

Data Curation

Organize, label, and annotate images and data to create analysis-ready datasets for reuse across your institution.

- Standardized research data model provides consistency, with extensibility options
- Data organized by groups, projects, and subjects with full support for longitudinal studies
- Add custom metadata to any object in Flywheel for maximum flexibility
- Integrated radiology-grade viewer with support for persistent, searchable ROIs
- Support for blind reader studies with customizable labels, annotations, and task assignment workflows
- Customizable forms in the viewer for clinical data capture, measurements, and workflow status

Automate and Compute

Automate and integrate image processing pipelines, machine learning workflows, and analytics

- Automate data pre-processing workflows and analytic pipelines via standard plug-in applications (Gears)
- Create custom Gears or utilize ready-to-use applications from the Flywheel Gear Exchange
- Container-based approach facilitates handling large batch processing workloads with Flywheel’s elastic compute scaling
- Open APIs and SDKs to enable rapid integration with existing tools and pipelines
- All capabilities of the Flywheel system are available via a published REST API and SDKs for leading scientific languages including Python, MATLAB, and R
- Rapidly select metadata and data for use in reports and analysis through Flywheel Views
- Choose data/metadata fields, review selected outputs, and export for manual external analysis
- Views can be saved and accessed via SDKs for use in tools such as Jupyter Notebook, R studio, and MATLAB
- Derivative information is recorded when Gears are run to support the consistency and reproducibility of your project, including version, inputs, status, and outputs
- Easy to ensure that all necessary processing steps were performed and performed consistently

Collaboration & Compliance

Empower secure collaboration and sharing while ensuring data privacy and regulatory compliance

- Secure, access-controlled projects allow for data and algorithm sharing with internal and external collaborators
- Manage privileges with a role-based permission model to ensure that only approved collaborators can access and modify data
- Federated identity service makes it easy to securely invite external collaborators to your Flywheel project
- Integrated with research access federations featuring 4,000 leading research institutions around the world to simplify user authentication. Federations include InCommon, EduGain, OrcID, the Australian Access Federation, and more
- Secure projects enable you to manage collaboration in compliance with IRB, HIPAA, and GDPR.
- Re-identification tools allow you to identify and purge data for research subjects if requested.
- Audit trails, digital signatures, and approval workflows ensure tracking and authorization. Additional options for 21 CFR Part 11 compliance.
- Provides radiologists with a lightweight, HIPAA-compliant medical imaging platform for developing algorithms for hospital workflows
- Researchers can run advanced models at scale on patient data and bring diagnostic analytics into patient treatment

Explore Flywheel’s Capabilities

Platform overview

Flywheel Demo

Accelerating Enterprise-Scale R&D

Schedule a Demo

Flywheel fits into any workflow

We’ve built a suite of products and integrations to deliver the best experience possible for our users.

Web Based UI

Flywheel’s easy-to-use web based interface is a powerful and robust system for working with and managing your medical research data.

Powerful SDKs

Flywheel offers SDKs (Software Development Kits) to suit your project’s needs including MatLab, R, and Python support.

Open APIs

Take advantage of our open source APIs to build your own tools and integrations, for truly infinite options for your research and machine learning development.

Flywheel Gear Exchange

Gears are ready-to-use plug-in applications that automate routine tasks, including metadata extraction, classification, quality assurance, format conversion, and full analytic pipelines. See over 70 plug-in algorithms available on the platform.

View Gear Exchange

Automate tasks with our open-source plug-ins

Health System Informatics Leader Uses Flywheel to Create AI-Ready Data Sets

Flywheel for Reader Studies: Demo

Streamlining Data Management for Multi-Site Traumatic Brain Injury Research

Flywheel Announces $54 Million in Series D Funding