1
Choose your source
LinkedIn or Facebook
Open this in another tab while logged into LinkedIn:
linkedin.com/mynetwork/invite-connect/connections/
linkedin.com/mynetwork/invite-connect/connections/
Open this in another tab while logged into Facebook:
facebook.com/friends
facebook.com/friends
2
Extract names & photos
Run the code in DevTools Console
- Go to your LinkedIn Connections tab
- Scroll to the bottom to load all connections (they lazy-load)
- Press F12 → Console tab
- Copy the code below, paste it into Console, press Enter
- An alert will say "Copied N connections" — come back here & paste in Step 3
⚠ LinkedIn lazy-loads photos. Scroll slowly through the full list before running the code,
or many profile pictures will be blank placeholder images.
3
Paste the extracted data
JSON from clipboard
4
Detect faces
Runs entirely in your browser — no data sent anywhere
TinyFaceDetector…
FaceRecognitionNet…
ℹ Face descriptors are computed locally using face-api.js (TinyFaceDetector + FaceRecognitionNet).
Nothing leaves your browser. Profile photos are fetched from LinkedIn/Facebook CDN — some will
be CORS-blocked; those people are saved as name-only entries and can still be imported.
5
Export & import
Choose how to get this into the face tracker
⬡ .glasseslib format v1
Both options export the same format. The library file is fully compatible
with the face tracker's ⬆ Import button — or push directly to
save_server.py if it's running locally.
⬇ Download File Always works
Downloads a .glasseslib JSON file to your computer.
Open the face tracker → click ⬆ Import → select this file.
⚡ Push to Server Local only
Instantly merges into the face tracker library via
save_server.py on localhost:8787.
No file needed — just run the server and click push.