// DATA CAPTURE & ANNOTATION STUDIO — BANJARMASIN

Human point-of-view footage,
labeled to train AI and robots.

We capture egocentric (POV) video in real-world settings — rivers, markets, city streets, and homes — and turn it into annotated datasets ready to train computer vision, robotic manipulation, and AI assistant models.

REC
00:00:00:00
POV_CAM_02
MARTAPURA RIVER
boat 0.94
person 0.97
boat 0.91
basket 0.88
10,000+
Hours of POV footage collected
75+
Annotated action & object categories
2
Task tracks: home & commercial
SERVICES

From the field to a training-ready dataset.

Our local capture operators film, our annotators label — everything checked through layered QA before a dataset ships.

01 · CAPTURE

Egocentric Video (POV)

First-person footage shot with wearable cameras, in real environments and activity scenarios you define — markets, kitchens, streets, workspaces.

02 · ANNOTATION

Labeling & Segmentation

Bounding boxes, object segmentation, action labeling, and hand/contact tagging — done by trained annotators following per-project guidelines.

03 · DATASET

Training-Ready Data for AI & Robots

Delivered in standard formats (COCO, custom JSON, and more) for training computer vision, robotic manipulation, or autonomous navigation models.

TASK TYPES

Two ways to work with us.

Whether you need footage from everyday home settings or a scoped commercial engagement, we run both as structured, QA-checked tasks.

HOME TASK

Household & Everyday Scenarios

POV capture of everyday home activity — cooking, cleaning, caregiving, hand-object interaction — recorded by our operators in real households. Ideal for training models that need to understand everyday human behavior.

COMMERCIAL TASK

Enterprise & Contract Projects

Scoped capture and annotation engagements for companies building AI or robotics products — custom scenarios, NDAs, larger volumes, and dedicated project timelines.

HOW WE WORK

Four stages, from empty tape to structured dataset.

00:00:01

Brief & Scenario

We build the capture scenario and annotation guidelines with your team, tailored to the model you're training.

00:00:02

Field Capture

Local operators shoot POV video on location — rivers, floating markets, homes, and city streets of Banjarmasin.

00:00:03

Annotation & QA

Annotators label frame by frame, then a separate quality-assurance pass verifies everything before it's finalized.

00:00:04

Deliver Dataset

The final dataset is delivered in the format you need, with documentation of its structure.

USED FOR

Datasets relevant across model types.

01

Robotic Manipulation

Hand-object data to train robots how to grasp and move objects.

02

Autonomous Navigation

POV footage of dense, dynamic environments for navigation & route-recognition models.

03

Home AI Assistants

Everyday activity from a first-person view, for models that need to understand household tasks.

04

AR / VR Training

Annotated POV datasets for models that need spatial and object-interaction understanding.

workspace 0.96
WHY US

Based in Banjarmasin, the city of a thousand rivers.

  • A genuinely varied capture environment

    The Martapura River, floating markets, waterfront settlements, and city streets — hard to replicate in a studio.

  • Trained local operators & annotators

    Recruited and trained directly by our team for consistent capture and labeling quality.

  • Flexible for custom scenarios

    Need a specific capture scenario for your model? We'll build it with you from scratch, home task or commercial.

START A PROJECT

Need an egocentric video dataset for your AI or robotics project?

Tell us about your scenario and data needs, home task or commercial — we'll reply with a capture plan and timeline estimate.

BANJARMASIN, SOUTH KALIMANTAN, INDONESIA