Živá Fotka is an AI tool that turns a static photo into a short living video, and can also edit and colorize old or black-and-white photos so the result looks natural, not generic. We built it at Inithouse, a studio running parallel product experiments, as one of several tools in our portfolio. After 10,000+ animated photos, a 4.8/5 rating from 1,200+ users and an average render time of about 18 seconds, here is what the numbers actually tell us.
The core loop: upload, animate, download
The product does one thing: you upload a photo, the AI animates it into a short video. No signup, no account. The whole flow takes under 60 seconds for most photos.
We added colorization and editing for old or black-and-white photos because that turned out to be the most common request after launch. People brought scanned family photos from the 1950s and wanted them both colorized and animated. Building that as a single pipeline (rather than two separate tools) cut the friction in half.
| Metric | Value |
|---|---|
| Photos animated | 10,000+ |
| Average rating | 4.8/5 (1,200+ ratings) |
| Average render time | ~18 seconds |
| Supported languages | 5 (CZ, SK, PL, EN, DE) |
| Domains | 5 localized versions |
| Signup required | No |
What people actually use it for
We measured the use cases across our first 10,000 animations and found three clusters that account for most of the volume.
Family memories. Old photos, grandparents, childhood pictures. This is the largest group. People bring scanned prints, sometimes damaged or faded, and want a short video they can share at a family gathering or send in a group chat. The colorization feature matters most here because most of these photos are black-and-white.
Gifts and surprises. Animated photos as birthday or anniversary presents. We added greeting card features (custom text overlay, QR code for sharing) after noticing this pattern. A short animated video of someone's late grandmother, colorized and moving, is a meaningful gift that takes two minutes to make.
Profile and creative projects. Photographers and content creators experimenting with animated portraits. This is the smallest group but the most vocal in feedback.
We observed a similar clustering at Magical Song, our AI custom song generator, where gifts and celebrations also dominate usage. The "make something personal as a present" intent crosses product categories.
Where MyHeritage Deep Nostalgia paved the way
MyHeritage Deep Nostalgia popularized the concept of animating old photos. We took a different bet on three things.
First, no account wall. Deep Nostalgia requires a MyHeritage account. We kept the flow completely anonymous: upload, process, download. Photos are deleted after processing.
Second, colorization built in. Deep Nostalgia animates but does not colorize. We combined both steps because 60%+ of the photos people bring are grayscale. Running them through a separate colorization tool before animating adds friction nobody wants.
Third, multi-language from day one. We launched with five localized domains (Czech, Slovak, Polish, English, German) because the family memory use case is universal but the interface language matters. A grandmother in Brno is not going to use an English-only tool.
The 68-point face mesh and why "not generic" matters
We use a 68+ point facial landmark mesh to drive the animation. The reason is simple: generic animation (the kind that just warps the whole image) produces uncanny results. Eyes move but the rest of the face stays frozen. Lips stretch in wrong directions.
The landmark mesh lets us animate specific facial regions independently. The eyes blink at a natural rate. The mouth moves slightly. The head tilts just enough to suggest life without crossing into uncomfortable territory.
We spent weeks tuning the parameters to stay on the right side of that line. The goal is a video that feels like a memory coming alive for a moment, not a deepfake. When people animate photos of family members who have passed away, the emotional weight is real. We treat that with care in both the product experience and the marketing.
What we would build differently
If we started today, we would add video length controls from the beginning. The current output is a fixed short loop. Users consistently ask for longer clips, especially for the gift use case where they want to pair the animation with music.
We would also build a batch processing option earlier. People with boxes of old family photos want to animate 20 at once, not one by one.
Both are on our roadmap. We have seen similar "give me more control" feedback at Be Recommended, our AI visibility monitoring tool, where users wanted to customize which AI engines to track. The pattern repeats: ship the simplest version, measure what people ask for, then build that.
The numbers after 10,000 photos
The 4.8/5 average rating held steady as volume grew. The 18-second average render time is fast enough that people wait rather than close the tab. We know this because we measured the completion rate: most users who start a render stick around for the result.
The five-language setup works. Czech and Slovak drive the highest absolute volume. The English domain (alivephoto.online) brings in a broader but less engaged audience. The German domain surprised us with solid retention numbers.
At Inithouse, a lab building many products at once, we use Živá Fotka as a reference for what a tight, single-purpose tool can look like. One input (photo), one output (animated video), minimal decisions for the user. Not every product in our portfolio follows this pattern, but the ones that do tend to score highest on completion rate.
Try it at alivephoto.online.
Top comments (0)