— Training Data Library
Task-specific datasets for robotics and physical AI.
Explore commercial, industrial, residential, and service task datasets designed for robot learning, perception, manipulation, and evaluation.
Manipulation
Pick and place, grasping, stacking, plugging, folding, assembly
Warehousing
Bin picking, packing, palletizing, scanning, inventory movement
Retail
Shelf scanning, restocking, product recognition, label reading
Assembly
Part insertion, fastening, fixture handling, sequence completion
Tool Use
Drills, scanners, grippers, hand tools, equipment operation
Service Tasks
Cleaning, delivery, object retrieval, hospitality workflows
Navigation
Indoor routes, obstacle interaction, scene understanding
Inspection
Defect detection, quality checks, visual verification
Human-Object Interaction
Grip, posture, intent, and sequence across everyday objects
— Compliance & Trust
Compliant by design, across every dataset.
From first consent form to final delivery, dexset workflows are built to meet data protection, security, and labor standards across the regions where we capture and the regions where our customers operate.
Informed consent
Every participant is briefed and signs a release before capture begins. Consent records attach to each clip, and withdrawal requests are honored across all delivered versions.
Privacy protection
Face blurring, anonymization, and exclusion zones are applied wherever required. We minimize personal data by design and never collect more than the task brief demands.
Data security
Footage is encrypted in transit and at rest, with role-based access controls, signed delivery URLs, and per-recipient transfer records on every dataset.
Regional compliance
Capture and delivery workflows are designed to align with GDPR and UK GDPR, CCPA/CPRA, PIPL, APPI, and LGPD requirements, with regional data-residency options where needed.
Traceability & audit
Source, consent, capture, environment, and annotation records persist for every dataset, supporting audits that run from raw footage to the exported training file.
Responsible sourcing
Capture operators, demonstrators, and annotators are fairly paid and work under documented, safe conditions — quality data should never come from exploitative labor.
— What Teams Say
Trusted by robotics teams.
★★★★★
“From brief to delivered COCO export in three weeks. Generic annotation vendors quoted us three months for lower quality.”
David Park
Founding Engineer, Independent Robotics Vendor
★★★★★
“We sent one task brief and got back a dataset that loaded into our pipeline on the first try. Custom schema, zero rework.”
Marcus Chen
ML Infrastructure Lead, Independent Robotics Vendor
★★★★★
“The egocentric hand-pose labels are the best we have evaluated. Our grasp success rate improved 18% after one fine-tuning round.”
Priya Raman
Manipulation Lead, AI Robotics Company
— FAQ
Frequently asked questions.
Are library datasets ready to use immediately?
Yes. Library datasets ship annotated, quality-scored, and documented, in standard formats. Most teams start with a library category to validate fit, then extend it with custom capture targeting their specific objects and environments.
Can library datasets be customized?
Yes. Any category can be extended with your objects, environments, camera setups, or label taxonomy. The library acts as a proven starting structure; custom capture fills the gaps your model actually has.
How are library datasets licensed?
Library datasets are licensed per organization for model training and evaluation. Custom extensions captured for you are delivered under exclusive terms. Licensing details, including redistribution limits, are agreed in writing before delivery.
What documentation comes with each dataset?
Every dataset ships with a datasheet covering capture context, label definitions, coverage statistics, quality scores, and known limitations — so your team can judge fit before a single training run.
Don't see your task?
Most of our datasets start as custom briefs. Tell us what your robot needs to learn.