One elevator · two people · 15 seconds · 16:9

I Love the Smiths AI

Upload two photos. One person wears the headphones. The other says the line. One unbroken take rebuilds the elevator scene from (500) Days of Summer — the glance, the headphone touch, the confession, the smile — in a 15-second horizontal video with the original sound mixed in.

No prompt to write. No editing. The casting order is the only direction that matters: photo 1 is the listener on the left, photo 2 is the speaker on the right.

Independent parody / fan project. Not affiliated with The Smiths, Searchlight Pictures, (500) Days of Summer, or any label. AI-generated — label it before you post.

1 Add two photos

0 / 2

One person per photo. The first is the listener with the headphones; the second is the one who says the line.

The listener

Stays original

The speaker

Stays original

Add two photos to start.

2 Pick the quality

16:9 horizontal. 720p is the native render; 1080p is an AI upscale of it. The 15-second original soundtrack is mixed in after the render.

This video: 16 credits · usually 5–15 minutes

One scene · one take

The whole bit, in one unbroken take

The model renders the full 15 seconds as a single continuous shot: no cuts, no stitched montage. The performance follows the reference frame for frame, and the original soundtrack is mixed back in after the render, so the lines land with the picture.

  • • One camera, one frame. The elevator, the wood panelling, the brass rail and the light never cut away.
  • • Two faces, one take. No cross-cut drift: the listener and the speaker keep the same identities for all 15 seconds.
  • • Original sound, added after. The scene's 15-second soundtrack is muxed in post; the performance is timed to it.
  • • 16:9 master. Post it wide, or crop to 9:16 in TikTok if you want the vertical feed.

Every order is generated from your two photos and the same continuous take.

Why this trend

A 2009 meet-cute became a 2026 template

The scene is from (500) Days of Summer: two people in an elevator, one wearing headphones, The Smiths playing so loudly that the other person can hear it. “I love the Smiths.” — “Sorry?” — “I said I love the Smiths.” It is a two-line joke about taste, timing, and misheard romance.

In 2024 the official clip resurfaced on TikTok. Through 2026 it became a couple/friend/coworker template: cast two people, keep the headphones, replay the lines, post it with the original sound. The hashtag has passed 12,000 posts and 300 million views; the “what I said” search variant is still climbing. AI versions arrived in November 2026 — most of them one long generation with soft cuts.

We render it as one continuous 15-second take, 16:9, with the original sound. No montage, no stitched cuts, no identity drift between shots.

How it works

Two photos in, one video out

01

Cast the two seats

Photo 1 = listener (left, headphones). Photo 2 = speaker (right, folders). Use clear, front-facing photos; a head-and-shoulders crop is enough. The order is the casting.

02

Pick a resolution

480p is the cheap draft, 720p is the default, 1080p is upscaled from the 720p master. All three are 16:9, 24 fps, 360 frames.

03

Render, download, post

One take renders, gets trimmed to 360 frames, and the 15-second soundtrack is mixed in. Download on desktop or scan the QR to save it on your phone.

Casting

Who goes where

The prompt names people by position, never by clothing. That is why the order is fixed and why you should not swap the photos to “see what happens”.

Listener seat example

1 · Listener

Left side. Wears the headphones. Stays in the elevator. Says “Wow.”

Speaker seat example

2 · Speaker

Right side. Holds the folders. Says the line. Reacts to the answer.

Photo rules

What makes a good upload

  • • One person per photo. No group shots, no mirrors, no heavy filters.
  • • Face clearly visible, eyes open, no sunglasses or mask.
  • • Head-and-shoulders or chest-up. Full-body photos are cropped automatically, but they lose detail.
  • • JPG, PNG or WebP, 100 KB–10 MB. Photos are deleted after 14 days.
  • • You must have permission from both people. No celebrities, no minors.

The render keeps each person's face and hair. The scene supplies the wardrobe — navy shirt and white headphones for the listener, cream cardigan and folders for the speaker — so the film look stays intact.

The file

Every parameter, before you pay

Duration15 seconds (360 frames at 24 fps)
Aspect / size16:9 horizontal · 1280 × 720 (720p), 854 × 480 (480p), 1920 × 1080 (1080p upscale)
Shots1 continuous take · 15.0s · no cuts
Audio15s original scene soundtrack, mixed after the render (no voice cloning)
InputExactly 2 photos · one listener, one speaker
WatermarkNone

16:9 is the master. Crop to 9:16 in TikTok or CapCut when you want the vertical feed.

Credits, not subscriptions

One video from $6.99

480p draft

7 credits

$6.99 pack · good for timing tests

720p default

16 credits

$12.99 pack · post-ready for TikTok

1080p upscale

26 credits

Best for YouTube / client work

Credits never expire. A failed render refunds automatically. Full pricing.

Compared

What the other generators give you

I Love the Smiths AIMakeViralilovethesmithsai.net
Length15s15s20s
Format16:99:16 (+16:9 full scene)16:9 / 9:16 / stacked
Photos221 or 2
Structureone continuous takeone multi-shot promptone 20s scene
Audioincludedincludedincluded
Watermarknonenonenone
Price (720p)$12.99credit pack$12.99

Public features as of November 2026. Not affiliated with any of the products above. Full comparison.

Questions

Before you cast it

All 25 answers

Your turn in the elevator

Two photos, 15 seconds, one take, no editing. If the render fails, you get the credits back.

Cast the two of you