Open Source FiftyOne

With over 5 million downloads, FiftyOne is the most popular open source visualizer and data curation tool for multimodal data.
Sync multi-rate sensor data to a shared timeline. Images, 3D scenes, plots, sensor streams, and metadata stay aligned as you scrub episodes.
Quickly scan thousands of samples or episodes on the grid.
Toggle labels, predictions, metadata, and custom fields to see exactly the information you need.
Organize and link multi-view samples with grouped datasets. Groups don't require a shared timeline.
Every visualization in Voxel51 is interactive, letting you inspect underlying samples with a single click.
Sync multi-rate sensor data to a shared timeline. Images, 3D scenes, plots, sensor streams, and metadata stay aligned as you scrub episodes.
Quickly scan thousands of samples or episodes on the grid.
Toggle labels, predictions, metadata, and custom fields to see exactly the information you need.
Organize and link multi-view samples with grouped datasets. Groups don't require a shared timeline.
Every visualization in Voxel51 is interactive, letting you inspect underlying samples with a single click.

FiftyOne Overview

Visualize, annotate, curate, and evaluate with FiftyOne.

FiftyOne lets you explore, curate, annotate, and evaluate datasets and models locally on your own machine.

Visualize

Explore multimodal episodes in context with synchronized playback across images, 3D scenes, sensor streams, maps, and annotations.
Multimodal data visualization in FiftyOne allows you to synchronize images, 3D, map, logs, plot, and message tiles on a timeline.

Annotate

Create, edit, review, and manage annotations across images, video, and 3D data. Inspect labels in context, fix inconsistencies, and prepare higher-quality training data.
Pedestrians annotated using SAM3 image segmentation

Curate

Compute and visualize embeddings to explore dataset structure, find similar samples, surface outliers, and uncover patterns that are hard to see with metadata alone.
UI screenshot of an "Embeddings" panel showing 48 colorful clustered data points with a lasso selection, and a labels menu listing "occluded_grasp," "handoff," "cluttered_bin," and "object_slip."

Evaluate models

Compare predictions to ground truth, inspect failure modes, and analyze model performance directly alongside the data your models were trained on.
UI screenshot of an "Agentic Labeling" workflow with a text prompt reading "Label frayed cables using the attached positive samples," four cable images with bounding-box annotations, a pipeline sidebar (Load dataset, Data selection, Agentic Labeling, HITL review, Annotate), and a "Run Agent" button.

OSS community

Loved by machine learning teams

0M+
product installs
0k+
GitHub stars
0+
enterprise customers

“FiftyOne is our primary resource for machine learning research. Thanks to FiftyOne's convenient field visualizations and filtering capabilities, we can easily distinguish incorrect labels and predictions, and therefore iterate on models faster than ever. As a result, we've achieved a 77% reduction in images sent for manual verification.”

Ryan Szeto
Senior Computer Vision Engineer at SafelyYou
Safely You logo in green

“As we developed our Florence-2 model, FiftyOne proved invaluable for data management and visualization. Its powerful capabilities helped streamline our workflow, ensuring we built a robust foundation for our models. Now, as we dive into the development of Florence-5B, we're relying on FiftyOne more than ever. The tool's intuitive interface and rich feature set are essential for effectively managing our large datasets and gaining critical insights.”

Bin Xiao
AI Researcher, Meta (formerly Principal Research Manager, Microsoft GenAI)
Microsoft logo

“FiftyOne has helped us speed up investigations by 3x. For example, if we see a wrong suction cup grasping an item, we can quickly visualize the issue across all data sources and identify what went wrong.”

Dimitry Pechyoni
Senior Principal Machine Learning Engineer at Berkshire Grey
Berkshire Grey logo

"FiftyOne enables researchers to analyze and improve the quality of their datasets rapidly, replacing the weeks of manual labor that would otherwise be required without this technology. High-quality data is critical to the success of machine learning systems. Without the right tools to analyze and curate datasets, machine learning development can be inefficient and ineffective.”

Jordi Pont-Tuset
Research Scientist at Google
Google logo

“At Allstate, my team works on auto vehicle damage inspection. Verifying the damage to a vehicle can take an insurance claim agent hours to verify, but using computer vision and FiftyOne, we can segment the parts of vehicles first, then detect the damages, and finally match the damage to repair costs and generate reports for the adjusters.”

Pavan Nanjundappa
Data Science Manager, Allstate India
Allstate logo

“We use FiftyOne to organize large research datasets. My favorite feature is the ability to view distributions over image attributes in the dataset, and filter the dataset by those attributes.”

Brett Israelsen
Principal Research Scientist, AI, Raytheon
RTX logo

“What really stands out about FiftyOne is the flexibility. The plugin framework lets us customize our workflows based on our unique needs, and the mature SDK lets us consolidate more of our pipeline into one tool, avoiding the cost of stitching together multiple systems. FiftyOne integrates directly into our production pipeline to drive 80% reductions in workplace incidents .”

Patrick Rowsome
Head of Computer Vision Operations, Protex AI
Protex AI logo

Voxel51 Upgrade

Take FiftyOne from local development to production

Upgrade to Voxel51 for the infrastructure, collaboration, governance, and automation teams need to run AI workflows in production.
Deploy anywhere
Diagram of FiftyOne Enterprise deployment options: cloud, hybrid, on-premise, air-gapped, and managed.
Fully customizable and extensible
Diagram of FiftyOne's customization options: front-end, security, plugins, and integrations, showing platform extensibility.
Support for billions of samples
FiftyOne dataset list showing image and video collections like Oxford Flowers 102 and Baseball Game with sample counts.
Dataset versioning
FiftyOne dataset versioning interface showing snapshots with sample counts and a rollback option.
Role-based access controls
FiftyOne interface showing role-based access controls, with teams assigned view or edit permissions.
Integrations
FiftyOne logo surrounded by integration icons: AWS, Anthropic, Google Cloud, OpenAI, Weights and Biases, Milvus, Hugging Face, NVIDIA, and MongoDB.

Customers

Trusted by enterprises

Why upgrade

Comparing open source FiftyOne and Voxel51

FiftyOne is the open-source project. Voxel51 extends OSS with governance, security, collaboration, and AI-powered workflows.
Overview
Curation, annotation, and evaluation
Users and data access
Single user, using local machine
Multi-user, with data in the cloud or on-prem
Deployment
Local machine
On-prem, air-gapped, cloud, hybrid
Data scalability
Limited by local resources
Scalable to 1B+ samples
Multi-user collaboration
—
Data governance & access control
—
Out-of-the-box workflows
Limited
Built-in and custom enterprise workflows
Workflow automation
—
Built-in scheduler for task delegation
Support
Community
Enterprise support with onboarding, upgrades, and maintenance
Details
Users
Users supported
Single
Multi-user
Deployment
Deployment options
Local machine
On-prem, air-gapped, cloud (private/public), hybrid, multiple environments.
Production and staging environments
—
Multiple
Dataset support and access
Multimodal data support
Data storage
Local filesystem
Cloud or on-prem media storage. Data lake integrations (e.g., Databricks, PostgreSQL)
Data volume
Limited by local resources
Cloud or on-prem media storage. Data lake integrations (e.g., Databricks, PostgreSQL)
Data scalability
Limited by local resources
Scalable to 1B+ samples with Data Lens
Cloud integration
—
Native integration with all major cloud providers (AWS/Govcloud, GCP, Azure, OCI, Tencent)
Team collaboration
Multi-user collaboration
—
Easy cross-team collaboration on datasets, views, individual samples, and model performance
Support
Support level
Community support
White glove enterprise support including dedicated communication channel, onboarding, upgrades, maintenance, and training.
Data governance & security
Data encryption
Users build their own security practices
Built-in support for encrypting sensitive data at rest and in transit
Data authentication and access
—
SSO integration with OIDC, OAuth2, SAML. User and role-based permissions
Dataset versioning and audit
—
Version datasets. Track, browse, and restore snapshots
Automation
Workflow automation
User implemented
Built-in scheduler to delegate tasks. Integrate with existing compute orchestration and resources.
Data workflows
Data visualization
Data embeddings
Manual data labeling
Auto data labeling
—
Auto-labeling, QA, and confidence scoring using foundation models.
Data quality
—
Built-in workflows with quality scoring to detect poor quality samples.
Data retrieval
—
Scenario search and retrieval at billion+ scale from data lakes using Data Lens.
Physical AI
—
Built-in workflows for data audit, augmentation, and neural reconstructions.
Model workflows
Model evaluation
Model comparison
Scenario evaluation
Sample-level analysis
Model eval execution
Manual
Built-in
Customization & Extensibility
Extensibility
Python SDK and plugins
Plugin and panel framework with support for custom enterprise workflows
Custom dashboards and analytics
—
Software licensing
License
Apache 2.0
Commercial license with enterprise terms

Join our FiftyOne community

Join more than 20,000 developers and scientists around the world who are leveling up their AI skills in
GenAI, Computer Vision, LLMs, MLOps, vector search, and more.
World map with orange markers showing FiftyOne's global AI, ML, and computer vision community.
Questions?
We've got answers.

Enough data wrangling.

Request a demo.