How it started
The first version of Percify was not an avatar tool at all. It pulled short videos apart to work out why they held attention: where the hook landed, how the beats were cut, what the captions were doing. That was genuinely useful, and it pointed somewhere better.
Because once you can read the structure of a video, you can perform it again. And the only thing missing is someone to perform it. It turned out a single photo was enough. The first Percify account was created on 30 August 2025 and the first avatar spoke on 1 October 2025. Everything since has been the same question asked in more places: what else becomes possible once there is a version of you that talks?
Four dates so far
A tool for taking videos apart
The first version read short videos to work out why they held attention. It was useful, and it pointed somewhere more interesting.
The first avatar spoke
If you can read the structure of a video you can perform it again, and one photo turned out to be enough to perform it. Voice cloning landed the same month.
Percify opened up
The public API and the MCP server went live, so Claude, Cursor and ChatGPT can make a video with your avatar without anyone opening the app.
One meter, everything on it
Avatars, voices, dubbing, viral remakes and 120+ hosted models, all drawing from a single credit balance.
Percify today
Counted in our own database, October 2026.
28,000+ accounts
People who have signed up since the first account, on 30 August 2025.
13,000+ avatars
Avatars built on the platform since the first one, on 1 October 2025.
180+ countries
Countries people have signed up from.
120+ models
Image, video and voice models you can run, all on one credit meter.
What we hold to
How the product is built, stated as things you can hold us to.
The product is you
Not a stock presenter reading your script. Percify starts from your face and your voice, and everything downstream keeps that identity intact.
One credit meter
Images, voices, videos and dubbing all draw from the same balance. Paid plans differ by monthly credits and API access, and by nothing else.
A free account, and a price before every run
Creating an account is free and needs no card. Making things needs a plan, and every model shows its credit cost before it runs.
Numbers with a date on them
Every figure on this site says what it counts and when it was counted. It keeps us honest, and it saves you having to ask.
The people behind it
A small team. These two started it.
Cypher
Started Percify in 2025 and leads the product: the avatar and voice pipeline, the studio around it, and what ships each week.
Drepheus
Runs the business side: operations, partnerships, billing and the day to day of keeping a platform dependable for the people paying for it.
What we are building next
Directions we are committed to, without dates we cannot keep.
Percify on iPhone
Your avatar in your pocket. Write or record a script on the phone and get the finished video back on it, no laptop in the loop.
Brand OS that measures
A brand already turns into a content plan. Next it reports back which of that content actually landed, campaign by campaign.
Flows, assembled end to end
Multi clip flows already plan and render every shot. The remaining piece stitches them into one finished video for you.
A deeper API
Public documentation, more tools on the MCP server, and volume pricing for the teams building products on top of us.
Want to talk to us?
The press kit has our brand files, the boilerplate and the current figures, all in one place. For anything else, including partnerships and integrations, [email protected] reaches us directly.
Questions, answered
What does Percify actually do?
It turns one photo of you into a photorealistic avatar that speaks, with your own cloned voice and lips in sync. From there it dubs video into other languages, remakes the structure of a short you paste in, and runs 120+ image, video and voice models on the same credit balance.
Do I need a camera or a studio?
No. One clear photo is enough to build the avatar, and a short sample of you talking is enough to clone the voice. After that you type a script, or record one, and the video comes back.
When did Percify start?
The first code was written on 24 August 2025, the first account was created on 30 August 2025, and the first avatar was generated on 1 October 2025. The public API and MCP server followed on 8 July 2026.
Who builds it?
A small team, founded by Cypher, who leads the product, and Drepheus, who runs operations and the business side.
Is Percify funded?
Percify is independent and has no outside funding round to announce. If that changes, the round and the investors will be named on this page.
How does pricing work?
One credit meter covers images, voices, videos and dubbing. Creating an account is free and needs no card; making things needs a plan, and plans differ by monthly credits and API access. There are no quality, speed or watermark tiers.
Are you hiring?
Not at the moment, and we would rather say so than list roles nobody can be hired into. If you build the kind of thing we build, [email protected] still reaches us.
