Interactive Samples & Dataset Specs
Browse 16 dedicated dataset categories below. Click any Category card to inspect hardware specs, quality standards, and standalone Type detail pages.
Humanoid Bimanual Teleoperation Data
Dual-arm robot movement datasets recorded from human operators for training humanoid AI hands.
High-frequency 50Hz dual-arm leader-follower teleoperation telemetry logging joint angles, velocities, and force-torque load vectors in HDF5 format.
Bimanual Kinematic Trajectories
50Hz joint angle and velocity time-series trajectory logs recorded from leader-follower arms.
Dual-Arm Haptic Force Load Cells
6-axis force-torque wrench load cell sensor feeds logging bilateral haptic feedback vectors.
Tactile Array & Slip Telemetry
High-density fingertip matrix pressure sensors recording object contact dynamics and micro-slip events.
Humanoid Bimanual Teleoperation Data
Dual-arm robot movement datasets recorded from human operators for training humanoid AI hands.
High-frequency 50Hz dual-arm leader-follower teleoperation telemetry logging joint angles, velocities, and force-torque load vectors in HDF5 format.
3D LiDAR Point Cloud & Spatial Cuboid Data
Laser point cloud scans with 3D bounding boxes for autonomous vehicles and spatial computing.
Calibrated 3D point cloud datasets with bounding cuboid annotations, point density classification, and camera-LiDAR sensor fusion alignment.
Egocentric First-Person POV Video Data
First-person perspective video captured from smart glasses for vision-language-action models.
4K 60fps first-person POV video capture from head-mounted EgoCAMs and binocular eye-tracking glasses for Vision-Language-Action (VLA) models.
Indic Multilingual Audio & Phonetic IPA Data
Speech voice recordings across 22 Indian languages for training voice assistants and ASR models.
Acoustic speech corpora recorded across 22+ Indic scheduled languages and regional code-switched dialects with time-aligned IPA phonetic tags.
Computer Vision & Data Annotation Services
Pixel-perfect image labeling, object segmentation masks, and skeletal pose keypoints.
Pixel-perfect bounding boxes, polygon semantic segmentation, 2D/3D keypoint landmark tracking, and multi-modal dataset annotation.
Human-in-the-Loop (HITL) & RLHF Data Studio
Human feedback ratings and safety prompt testing for aligning foundation AI models.
Domain-expert preference ranking, RLHF alignment data, adversarial red-teaming prompts, and expert human evaluation for foundation models.
Process Supervision & Chain-of-Thought Data
Step-by-step logic and math reasoning step checks to train reasoning AI models.
Step-by-step reasoning verification datasets evaluating intermediate logic steps for complex math, coding, and multi-step AI reasoning.
Procedural Synthetic Data Generation
Physically accurate 3D generated images and scenes for training AI without privacy risks.
Physically accurate synthetic imagery, procedurally generated 3D scenes, and edge-case simulation datasets paired with real-world ground truth.
Digital Twins & Physics Simulation Rigs
Simulated 3D factory environments for testing and training robots before real-world deployment.
Calibrated 3D digital twin assets, URDF/MJCF kinematic models, and physics simulation environments for NVIDIA Isaac and Mujoco simulators.
Robotic Teleoperation Pipeline & Hardware Rigs
Hardware motion capture rigs and haptic feedback systems for remote robot control.
End-to-end leader-follower teleoperation infrastructure, haptic exoskeletons, and data recording clients deployed in studio and field environments.
Domain-Expert Legal & Tax RLHF Studio
Specialized legal and tax reasoning data verified by certified lawyers and accountants.
Vetted lawyers, chartered accountants, and tax specialists providing high-precision RLHF ratings, instruction tuning, and compliance audits for LLMs.
Localized Dialect & Conversational Audio Bank
Authentic field recordings of regional Indian accents and spoken conversational dialects.
Multi-region field speech recordings across 22+ Indic scheduled languages, regional dialects, and authentic code-switched conversational audio.
Synthetic-to-Real (Sim2Real) Vision Pipeline
Paired synthetic 3D and real camera imagery designed to bridge the real-world gap.
Paired synthetic and real-world image datasets designed to measure, analyze, and close the sim-to-real domain gap in computer vision models.
Surgical & Medical Vision Annotation Bureau
Sub-millimeter accurate medical vision labeling for robotic surgery and diagnostic AI.
Sub-millimeter accurate medical vision annotation, surgical instrument tracking, tissue segmentation, and DICOM dataset processing.
360° SLAM Scanners & Mobile Reality Capture
360-degree spatial video scanning and indoor 3D mapping for facility digital twins.
360° spatial video scanning, mobile LiDAR SLAM trajectory logging, and dense 3D mesh reconstruction for indoor facility mapping.
3D Output Formats & Dataset Schema Guide
Standardized dataset specs and converters for HDF5, OpenUSD, Parquet, and Point Cloud files.
Comprehensive guide and converter tools for production dataset formats — HDF5, Parquet, ROSbag, Open X-Embodiment RLDS, COCO JSON, and PCD.
Off-The-Shelf (OTS) Physical AI & Robotics Datasets
Pre-cleared, production-scale datasets with complete hour statistics, sensor calibration parameters, and direct Google Drive task sample links below. Includes 150,000 hrs Stereoscopic 3D Dual-Camera, 300,000 hrs First-Person Monocular, and HMD 6-DOF IMU Telemetry across 30+ hubs worldwide.
Stereo Camera 3D Video • Task & Resolution Breakdown
Synchronized binocular feeds with hardware genlock, calibrated baseline geometry, and stereo depth. Click any task sample below to inspect verified Google Drive folders.
Stereo 2560×720 Dual-Sensor HD
Dual synchronized global shutter streams with 60mm baseline geometry. Ideal for fine-motor hand manipulation, spatial disparity estimation, and robotic grasping.
Stereo 3840×1080 Dual Full-HD
Side-by-side Dual 1080p high-resolution stereo capture with calibrated extrinsic transformation matrix. Engineered for humanoid navigation and outdoor walking.
Stereo 50+ Stream Library & Specs
Full archive of 50+ diverse continuous stereo video streams paired with hardware specification sheet and calibrated octet sensor telemetry.
Task-Specific Breakdown • Direct Google Drive Sample Links
Each task contains time-aligned 4K video clips, head-mounted 6-DOF IMU trajectories (accelerometer + gyroscope), and fine action boundary timestamps. Click any task to inspect samples:
Cooking & Food Prep
Vegetable chopping, food processing, stove pan stirring, coffee brewing, ingredient weighing, multi-utensil food staging.
Dishes & Kitchen Cleanup
Table clearing, sponge washing, plate stacking, dishwasher rack loading, countertop sanitation, drying rack placement.
House Cleaning & Sanitizing
Vacuuming carpets, microfiber surface wiping, window mirror polishing, broom floor sweeping, bathroom tile scrub.
Laundry & Garment Care
Color sorting, washing drum loading, detergent measurement, dryer cycles, garment folding, hanger arrangement in closet.
Groceries & Pantry Staging
Cart unboxing, carton stacking, barcode identification, cold fridge sorting, spice rack alignment, shelf stock replenishment.
Car Cleaning & Wax Buffing
Applying wax, rotary buffing polish, car upholstery vacuuming, windshield mirror streak-free wiping, wheel rim degreasing.
Tidying & Spatial Organizing
Decluttering desks, sorting cables and chargers, returning books to shelves, organizing storage boxes, cushion alignment.
Outdoor & Yard Care
Patio swept-clean routines, rake lawn leaves, carry garden tools, hose watering, deck furniture arrangement.
Plant & Pet Care
Watering can soil dosing, plant leaf misting, pet food dish filling, grooming brush strokes, toy cleanup.
First-Person Monocular Industrial POV Library
Raw 4K head-mounted egocentric video captured across operational industrial plants, factories, and agricultural lands.
Industry Task Taxonomies & Sample Stream Explorer
Explore real-world task procedures across 10 specialized industrial domains. Switch tabs below to preview sample clips, multi-angle camera perspectives, and annotation specifications.
🧵 Garments, Textiles & Power Loom Weaving
Industrial sewing machine stitching, overlocking, cloth cutting, pattern grading, loom operation, steam pressing, and final packaging. Captures non-rigid material contact dynamics and hand dexterity.
Interactive Samples & Dataset Specs
Explore sample downloads, sensor schemas, and video streams across robotics teleoperation, spatial computing, egocentric vision, and Indic speech.
Looking for Custom Sensor Configurations?
Tell us your target joint angles, frame rates, or dialect matrices and access our live Google Drive repository.
Upcoming & Specialized Industry Verticals
Beyond our core robotics and multimodal foundation data lines, we mobilize turnkey sensor rigs and managed field annotators across 8 specialized industry sectors:
Autonomous Vehicles & In-Cabin AI
850/940nm NIR driver eye-gaze tracking, micro-sleep (PERCLOS), child seat occupancy (OMS), 128-beam LiDAR, and complex unstructured Indian traffic edge cases.
Precision Agriculture & AgriTech AI
Sub-pixel weed vs. crop segmentation for laser weeders, disease/pest pathology macro datasets across 60+ crops, and 5-band drone NDVI/NDRE canopy mapping.
Smart Manufacturing & AOI Inspection
Automated Optical Inspection (AOI) microscopic scratch/burr/solder defect segmentation, 192kHz acoustic bearing failure audio, and worker PPE compliance.
Drones, Defense & Geospatial GIS
ASPRS classified aerial LiDAR (DTM/DEM), 0.3m GSD satellite building/road polygons, 8–14µm thermal night ISR camouflage tracking, and GPS-denied UAV odometry.
Smart Retail & Automated Checkout
10,000 sq ft mock supermarket testbed: Overhead fisheye shopper Re-ID tracking, 60fps shelf item pick/put-back action boundaries, and self-checkout theft AI.
AR / VR & 3D Biometrics
468-point 3D facial meshes, 52 ARKit BlendShapes, headset-view 21-keypoint 3D MANO pinch tracking, 24-camera full-body MoCap, and face liveness (PAD).
Fintech & Insurance Fraud AI
Handwritten document OCR across 22 Indic scripts, paired signature forgery detection, 4K vehicle collision damage part segmentation, and CA/CPA tax reasoning PRMs.
Renewable Energy & Infrastructure
Radiometric thermal solar PV inspection (IEC 62446-3), 5-tier wind turbine blade leading edge erosion, transmission grid LiDAR vegetation clearing, and pipeline corrosion.
Speech Intelligence & Conversational AI
Full-duplex 32-bit float conversational audio with natural turn-taking, overlapping speech, 22 Indic languages, 85+ dialects, and domain-specific clinical/legal/banking dictation.