How to Make an I Love The Smiths AI Video With Two Photos

To make an I Love The Smiths AI video with your own photos, prepare one clear image of each consenting adult, assign Photo 1 to the headphone listener and Photo 2 to the speaker, then use a generator with the dedicated elevator-scene reference. This site builds the editing prompt for you and is designed to preserve the reference timing and sound. Its current release supports free setup testing, but live generation is disabled.

Before you start

You need:

• Two separate photos: JPG, PNG or WEBP, up to 10 MB each;
• One consenting adult per photo: no group photos, minors or public figures;
• A clear view of each face: sharp focus, even lighting and no sunglasses, mask or heavy filter; and
• Permission for both likenesses: use your own image or obtain explicit agreement.

No prompt, editing timeline or elevator footage is required.

1. Watch the supplied reference

Check the two roles, camera, dialogue and timing before choosing photos. The supplied MP4 is 1280 × 720 and approximately 18.901 seconds long. The editing request is designed to retain that source duration rather than substitute a fixed 12- or 15-second template.

2. Choose and assign both photos

Upload Photo 1 for the headphone listener, who stands on the left and stays in the elevator. Upload Photo 2 for the conversation starter, who says the line and walks out. Upload order matters: swapping the files swaps the assigned roles.

Use a natural head-and-shoulders or waist-up portrait with one face clearly visible. Different clothing colors can make the two people easier to distinguish, but identity and motion can still vary in generated results.

3. Choose a resolution and check the setup

The configured cost is 50 credits for 720p or 75 credits for 1080p. Resolution changes the requested output quality, not the reference-duration instruction.

During the current development release, the preview button validates the two-photo setup without submitting a model request. Selecting photos or testing this setup does not generate a video or consume AI credits.

4. Generate and recover the task when live access is enabled

Once live generation is separately enabled, the server will upload the inputs, run safety checks, create an AI task and submit the fal.ai request. Generation requires an account and enough credits. Provider latency varies, and the task list is designed to restore active work after a refresh.

A confirmed failed generation returns the credits consumed by that task. Payment confirmation and credit grants are handled server-side; returning from checkout alone does not grant credits.

5. Review the MP4 before downloading or posting

Watch the entire result rather than judging it from the first frame. Check both identities, face and hand motion, lip synchronization, dialogue, sound and scene continuity. Download the MP4 only after you are satisfied, and label it as an AI edit when sharing.

No verified project-generated output is published yet. Template demonstration media is not a substitute for reviewing your own result. When live access is enabled, check the actual output duration and quality instead of assuming that the duration-preservation instruction guarantees an exact match.

How do I post it as an I Love The Smiths TikTok video?

When generation is enabled and your MP4 is ready, import it into TikTok, Instagram Reels or YouTube Shorts and check that both faces remain inside the platform’s crop. The current workflow does not expose a separate aspect-ratio control or guarantee a platform-ready 9:16 result, so reframe it in an editor if necessary.

Do not assume that a reference soundtrack grants reuse rights on every platform. Use the platform’s licensed sound tools when needed, keep the AI label visible and follow each platform’s rules.

For cost and availability details, read is there a free I Love The Smiths AI generator? For background and terminology, see what is the I Love The Smiths AI video trend?