Instagram Multi-cam Stories: cut phone and glasses angles on one audio

Instagram's Multi-cam format "merges footage from glasses and phone, with automatic clip synchronization", per the Instagram for Creators post on AI glasses, dated June 30, 2026. Sume does not detect sync. What it can do is cut between two angles of the same moment when you tell it the offsets: one audio spine, several video[] slots, each with a source_in. That is a manual version of the same idea, suited to an angle switch you already planned.

What does Instagram's glasses toolkit include?

The post covers Ray-Ban Meta, Oakley Meta and Meta Glasses. You upload a Story from the gallery, pick glasses media (marked with a glasses icon) and tap the glasses icon on the edit screen to reach the new tools.

How do I cut two angles over one audio track?

Pick the angle with the better sound and detach its audio once with audio detach (wav by default, output up to 900 seconds, $0.01 per job). Use that as audio.url, the spine. Then declare slots that alternate between the two videos. Because the spine runs on one clock, set each slot's source_in to the same time in that angle's own file. If the two files do not start at the same instant, add the offset you measured, since Sume will not find it for you.

What does the plan look like?

This script builds a 20 second cut that switches angle every 5 seconds, where the glasses file starts 0.8 seconds after the phone file. It prints a body for the unbilled plan route. A slot needs a duration of at least 0.2 seconds, video[0].start must be 0, and later starts must increase.

import json

phone = "https://media.sume.com/artifacts/artf_demo/phone.mp4"
glasses = "https://media.sume.com/artifacts/artf_demo/glasses.mp4"
glasses_offset = 0.8   # seconds the glasses file starts after the phone file
slots = []
for i in range(4):
    t = i * 5
    if i % 2 == 0:
        slot = {"source_url": phone, "start": t, "duration": 5, "source_in": t}
    else:
        slot = {"source_url": glasses, "start": t, "duration": 5, "source_in": t - glasses_offset}
    slots.append(slot)

body = {
    "audio": {"url": "https://media.sume.com/artifacts/artf_demo/phone.wav", "duration_seconds": 20},
    "video": slots,
}
print(json.dumps(body))

What can go wrong?

The most common problem is drift: if you guess the offset, lips and sound disagree at the switch. Render the plan and watch the first switch. A hard cut is allowed; a transition is optional, up to 1 second, and more than 8 chained fades are refused. If you need Instagram's automatic sync, use it in the app, and use Sume for the parts around it such as a voice-over spine or a longer edit.

Is this only for glasses?

No. Any pair of angles of the same event works the same way: two phones at a talk, or a screen recording and a face camera. Meta's Multi-cam is specific to glasses plus phone footage; the Sume approach is just ordered slots on a shared clock.