backed by GIX Robotics Labs

egomic.

the industrial sop data layer for embodied ai.
contact us
trusted by 3 of the world's frontier ai labs
our offering

three pillars of industrial intelligence.

01

data

the most diverse industrial sop dataset across automotive, electronics, garment, seafood, batik & metalwork — captured first in indonesia, then southeast asia.

02

tech

a temporal vlm pipeline (qwen2.5-vl + hamer + sam2 + foundationpose) that turns raw factory video into structured sop steps, hand keypoints and 6dof object poses.

03

network

boots on the floor in 8+ industries, with consent forms, anonymization, and continuous qa. factory owns its data; we license a clean, anonymized cut.

data

egocentric video data captured in the real world.

first-person footage from indonesian factories with sop-level process knowledge — not staged demos, not lab data.

laboratory▶ playing
electronics
automotive
garment
batik
seafood
pharma
metal fab
palm oil
rubber
jewelry
furniture
pottery
shoes
leather
spice
coffee
chemicals
agriculture
plastics
glass
laboratory▶ playing
electronics
automotive
garment
batik
seafood
pharma
metal fab
palm oil
rubber
jewelry
furniture
pottery
shoes
leather
spice
coffee
chemicals
agriculture
plastics
glass
+ many more
tech

a temporal vlm that turns raw factory footage into structured knowledge.

training within industry (twi) sops paired with egocentric footage. each task broken into steps, key points and reasoning — auto-annotated, human-verified.

now showing: laboratory← select a tile above to change
LIVE PREVIEW
0:00.0
loading annotations…
// Annotation Sequence0 steps
// no annotations loaded
click a step to seek
network

a growing network of factories across indonesia & southeast asia.

sops captured across 8 industries with continuous qa, worker consent and clean data rights — owned by the factory, licensed to us.

0
factories
0
countries

get in touch.

whether you're building foundation models, deploying industrial ai, or running a factory in southeast asia — let's talk.

contact us
→ open sop annotation viewer