How to Make the Coffee Fail AI Video: People, Dog and Cat Versions From Your Photos

Two cooks behind a kitchen counter, one cup of coffee to make, and nothing goes right: grounds spooned straight onto the counter, hot water into an empty mug, a whole jug of milk across the worktop, and not a flicker on anyone's face. That's the coffee fail, the AI kitchen gag that took off in late October 2026 with two dogs and was soon remade with cats and people. Here's how to make it from your own photos.

What is the coffee fail video?

It's a 15-second AI clip filmed in one take from a phone propped on the counter. The one in front spoons coffee grounds from the jar onto the counter next to a red mug, again and again, deadpan. The one behind pours hot water into the empty mug, then tips a jug of milk straight onto the counter. A shared grin, a thumbs up, and the mug lifted with a straight face. The song is Julio Iglesias's "Con la Misma Piedra".

Our Coffee Fail template follows that take — the moves, the camera and the timing — in three versions: two people, two dogs or two cats.

How to make it

1. Open the Coffee Fail template.
2. Pick the version: two people, two dogs or two cats.
3. Add one or two photos. Photo 1 makes the coffee in front, photo 2 pours the milk behind. One face or one animal per photo.
4. Generate. Every version renders the whole 15 seconds.
5. Download the MP4. It comes with the song. To pick the sound yourself when you post, use Download without music instead.
6. Post it to TikTok, Reels or Shorts.

What happens in the video

• 0–4.5 s: grounds on the counter. Photo 1 spoons coffee next to the mug, not into it; photo 2 pours hot water into the empty mug.
• 4.5–9 s: more spoonfuls. Another, and another, onto the same pile. Total deadpan.
• 9–12.25 s: the grin. The two catch each other's eye and crack up.
• 12.25–13.25 s: the milk flood. Photo 2 empties a jug of milk across the counter; photo 1 gives a thumbs up (a raised paw for pets).
• 13.25–15 s: the sip. Photo 1 lifts the mug of plain hot water with a perfectly straight face.

Only one photo?

That works too. Add just photo 1 and it makes the coffee; the AI adds an ordinary partner to pour the milk — a made-up person in the people version, a golden retriever in the dog version, a ginger cat in the cat version.

Photos that work best

• Face clearly visible. Photo 1 is close to the lens and grins straight at it, so its face matters most.
• Head and shoulders are enough. Both cooks are seen from the waist up behind the counter. For pets, show the ears and the whole head.
• Good light. Daylight or a bright room; a dark photo makes faces and markings drift.
• One face or animal per photo. No second face, no person holding the pet.

A dog and a cat together? Use the dog or the cat version and add one photo of each: each animal keeps its own species, face and markings.

What does it cost?

Premium is 30 credits at 480p, 60 at 720p and 105 at 1080p; Balanced starts at 15 credits. Without an account, one video costs from $7.99 at checkout. No watermark, and failed renders and rejected photos are refunded automatically.

FAQ

Do we need to be in a kitchen in the photos? No. Any clear photo works. The kitchen, the counter, the mug, the kettle and the milk are generated.

Is the song included? Yes. Your video plays and downloads with "Con la Misma Piedra". Or download it silent and add the sound in TikTok.

How long does it take? Usually a few minutes. Your photos are reviewed first, then the video renders.

Make your Coffee Fail video →