Who is this again?
The person you met at the meetup shows up again. The lens shows their name, where you know them from and what is still pending.
VisualContext · testing on real hardware
Your glasses already see. VisualContext makes them remember.
You know the face, but the name will not come. Nor where you met, nor what you agreed on. VisualContext keeps that context and brings it back in the corner of the lens, the moment the person shows up.
Short sessions, nothing to install.
Through the lens
Three moments when memory fails.
Everything shows at the top of the lens, in green, and goes away when you no longer need it.
Save it or let it go.
A conversation that matters becomes a contact after a short countdown. A trivial one, like with the gas station attendant, you drop with one tap and it does not insist.
Where are my keys?
VisualContext remembers where it last saw your keys, wallet, car or motorbike, with a photo of the spot.
Manager · phone and web · AI search in development
Ask the way you would ask a friend.
The dashboard keeps people, places, things and memories. Instead of filters and forms, you ask in plain language, and every answer shows where it came from.
Preview with demo data. AI search is in development; the dashboard card is the same one that shows on the lens.
How it works
Encounter, record, recall.
- 01
Encounter
The glasses notice the person or object in front of you.
- 02
Record
What matters becomes text: name, context and what was said. Audio and video are not kept.
- 03
Recall
Next time, the context comes back on the lens. On your phone, you ask and correct by voice.
Privacy
Your memory, not surveillance.
Only people you saved
Nobody becomes a contact without your confirmation. What you drop is gone.
Text, not recordings
No raw audio or video is kept. The only visual data is a small crop of the face.
Limited transcription
Transcription only happens when a saved person is in the conversation.
Nothing leaves on its own
VisualContext does not share data with other apps or agents by itself.
Settings
On the glassesFacesRecognizes who is in front of you and records the encounter
PlacesCards pinned to nearby buildings
ThingsMotorbike, keys and whatever you ask it to remember
MemoriesAsk your memory through the glasses
Starts recording when
PrivacySomeone speaksA look is enough
Looking without talking does not record. The start of a conversation is never lost: the last 30 s stay only in the device memory.Protection in sensitive situationsOn by default›
How long to keepTranscripts: 1 year · faces: until you delete them›
Forget everythingErases memory and keys from this device›
Manifesto
Recognizing things has become a commodity: reading text, describing the scene, translating.
What is missing is remembering. Whose face that is, what you talked about, where the keys ended up.
VisualContext does not act for you. It records, and brings back the context at the right moment.
An extension of memory, for anyone who has ever forgotten a name they should have known.
Marcus de Magalhães, founder of alema.io
Want to try it on your glasses?
VisualContext runs on Rokid Glasses and is in testing. Write to join the beta list.