See through human eyes.
First-person video that captures how people navigate spaces, handle objects, and complete real tasks.
Explore egocentric data ↗ THE FOUNDATION FOR EMBODIED AI
Real-world data.
Extraordinary
intelligence.
The next leap in AI starts with the world around us. Give your models the human perspective, precise actions, and real-world experience they need to move forward.
+ FROM HUMAN EXPERIENCE TO MACHINE CAPABILITY
HUMAN PERSPECTIVE / ROBOT ACTION / INTELLIGENT SYSTEMSDiscover what’s possible ↓
REAL EXPERIENCE.
AT MEANINGFUL SCALE.
AT MEANINGFUL SCALE.
120k+
Hours of ego & G1/G2 robot data
10k+
Bimanual UMI episodes
08
Specialist dataset collections
01 / THE ACTIONABLE ADVANTAGEIntelligence needs
Intelligence needs
experience.
Models learn the world through data. We help you find the right perspective, the right actions, and the right context to turn that learning into capability.
Connect vision to action.
Bimanual demonstrations pair visual observations with gripper pose and action for manipulation learning.
Explore UMI demonstrations ↗Learn from real robots.
Multi-view recordings, state, and action data from physical robots performing everyday manipulation.
Explore robot datasets ↗02 / THE DATA CATALOGUEYour next breakthrough
Your next breakthrough
starts here.
Explore eight collections across human observation and robot action. Find the data that fits your model, your task, and your ambition.
8 collections to exploreSelect up to 3 datasets to compare
EGOCENTRIC / STEREOAA / E01Stereo Ego · 1080p
Human perspective, captured in stereo with audio and motion sensing.
3,400 hours
Dual 1080p30 fpsMCAP
EGOCENTRIC / STEREOAA / E02Stereo Ego · Wide-view
A wider perspective on complete manufacturing and craft workflows.
63,060 hours
190° field of view30 fpsMCAP
EGOCENTRIC / MONOCULARAA / E03Monocular Ego · Android
Everyday skills and workplace tasks, seen from a human point of view.
32,884 hours
1080p + portrait30 fpsMP4
EGOCENTRIC / MONOCULARAA / E04Monocular Ego · iPhone
Continuous first-person recordings that follow activities from start to finish.
9,832 hours
Up to 60 fpsMP4 / MOVContinuous tasks
MANIPULATION / UMIAA / U01Bimanual UMI
Connect human demonstrations to precise gripper movement and action.
10,000+ episodes
3 camera views60 Hz actions6DoF pose
ROBOT / MULTI-VIEWAA / R01Real Robot · G1 / G2
Whole-body and dual-arm experience from robots in real environments.
12,000 hours
8–9 camerasLeRobot v3.0State + action
ROBOT / MANIPULATIONAA / R02Galbot Robot Data
Multi-view robot demonstrations across household manipulation tasks.
14,997.91 QC-passed hours
4 RGB views30 fpsMCAP
ROBOT / MULTI-STEPAA / R03Long-Horizon Robot Data
Multi-step, bimanual demonstrations for learning sequences of action.
Let’s scope your project
RGB + depth250 Hz arm actionsJSONL
Some breakthroughs
need a different
kind of data.
Let’s build a collection around your tasks, your environment, and your ambitions.
Scope your dataset ↗YOUR NEXT MOVE STARTS HERE.No matching datasets.
Try a different task, format, or keyword.
Your task. Your environment. Your dataset.
Need a specific scene, skill, or capture setup? Let’s scope a custom collection.
03 / BUILT FOR WHAT’S NEXTFrom observing the world
From observing the world
to acting in it.
Different models. Different challenges.
A common need for meaningful real-world experience.
Vision & world models
Rich first-person context for visual understanding, activity recognition, and learning how the physical world changes.
Robot learning & VLA
Visual observations paired with movement and action, for imitation learning and vision-language-action research.
Long-horizon skills
Multi-step demonstrations for studying task planning, bimanual coordination, and sequences of real-world actions.
04 / FROM EXPLORATION TO EXECUTIONA clear path to
Start a conversation ↗A clear path to
the right data.
Tell us what you’re building.
Share your model, target tasks, and data requirements. We’ll help define the collection that fits.
Evaluate the fit.
Review sample access, capture specifications, annotation, and package structure with our team.
Move forward with clarity.
Agree the scope, licensing, and delivery plan. Start with existing inventory or build a custom collection.
LET’S MAKE IT ACTIONABLEBig ambitions.
Big ambitions.
Meet real-world
data.
Tell us what you’re building. We’ll help you find the data to take it further.
info@actionable.ai ↗Based in Singapore. Built for what’s next.68 Circular Road #02-01
049422, Singapore
049422, Singapore