Voxel51
Voxel51 brings transparency and clarity to the world's data by helping AI builders search, visualize, curate, label, and evaluate multimodal data. Its platform turns raw recordings into high-quality training and evaluation data for physical AI and visual AI models.
At a Glance
- Enterprise AI/ML teams
- Automotive and autonomous-vehicle companies
- Robotics companies
- Manufacturing
- +6 more
AI Tools by Voxel51
(1)
FiftyOne
Computer Vision Dataset Tool
Discussions
No discussions yet
Be the first to start a discussion about Voxel51
Latest News
Talk to Your Dataset: A Query Agent for FiftyOne, powered by Laguna S 2.1
Introduced Voxel51 Search for multimodal physical-AI data discovery
Introduced multimodal time-series support for VLA and physical-AI workflows, including MCAP ingestion, synchronized visualization, temporal tagging, and embeddings
Showcased FiftyOne, NVIDIA Omniverse simulation-to-reality workflows, zero-shot auto-labeling, and Twelve Labs/Databricks video workflows at CVPR 2025
Products & Services
Apache-2.0-licensed developer toolkit for building high-quality datasets and visual AI models; supports visualization, dataset curation, annotation, model evaluation, plugins, and integrations.
Managed enterprise version for secure collaboration on cloud-hosted datasets and models, dataset search, sharing, and version management.
Commercial platform adding multiuser collaboration, cloud data support, on-demand compute, enterprise security, deployment options, and support; branded Voxel51 as of September 2026.
Open-source AI/ML capabilities for the FiftyOne ecosystem, including embeddings, similarity search, uniqueness, and label-quality analysis.
Market Position
Voxel51 positions itself as an end-to-end multimodal data platform for physical AI, differentiated from labeling-only or model-training tools by combining open-source FiftyOne with enterprise collaboration, dataset search, curation, annotation QA, model evaluation, and synchronized multimodal/3D workflows. Comparable data-centric AI and computer-vision platforms include Encord, Scale AI, Labelbox, V7, Lightly, and Roboflow; FiftyOne's open-source ecosystem is a key differentiation.
Leadership
Founders
Jason Corso
University of Michigan professor (Toyota Professor of AI) and computer-vision/robotics researcher; co-founded Voxel51 with his PhD student Brian Moore.
Brian Moore
Computer-vision and machine-learning researcher who was Jason Corso's PhD student at the University of Michigan; co-founder and CEO of Voxel51.
Executive Team
Brian Moore
Co-Founder and Chief Executive Officer
University of Michigan computer-vision and machine-learning researcher; led Voxel51 strategy and vision after co-founding it with Jason Corso.
Jason Corso
Co-Founder and Chief Science Officer
University of Michigan Toyota Professor of AI and computer-vision/robotics researcher; leads the company's scientific direction and FiftyOne ecosystem.
Board of Directors
Founding Story
Voxel51 was founded at the University of Michigan in 2018 when Professor Jason Corso and PhD student Brian Moore turned their research into developer-friendly tooling for computer-vision data. They created FiftyOne to address the difficulty of understanding, curating, and improving real-world visual datasets and models.
Business Model
Revenue Model
Commercial subscriptions and enterprise licensing for managed Voxel51/FiftyOne deployments, priced by seats and compute capacity (VPUs), with custom enterprise services and support; FiftyOne Open Source is free under Apache 2.0.
Pricing Tiers
8 user seats, 16 guest seats, 4 VPUs, 2,800 hours/month compute, one production deployment, one dev/staging deployment, unlimited data and model inference, SSO, standard enterprise support.
25 user seats, 100 guest seats, 20 VPUs, 14,000 hours/month compute, three production and three dev/staging deployments; adds on-premises/air-gapped deployments, premium support, and a dedicated customer-success engineer.
Unlimited users, guests, VPUs, and deployments; adds a dedicated customer-success team and professional services.
Target Markets
- Enterprise AI/ML teams
- Automotive and autonomous-vehicle companies
- Robotics companies
- Manufacturing
- Healthcare and medical imaging
- Defense and government
- Physical AI and robotics
- Autonomous vehicles and advanced driver assistance
- Manufacturing and visual inspection
- Healthcare and medical imaging
- Defense, aerospace, and security
- Retail and logistics
- Walmart
- Honda
- Michigan Medicine
- NVIDIA