Data
Match and season numbers from football data services, plus your club's own numbers from your statisticians and the organization's.
- Internal app
- Spreadsheet
- Paper
- Voice note
Soccer analytics for clubs, federations and organizations
FF, our AI, reads your season. Match data, your staff's statistics, GPS and video, joined on one timeline: FF shows it in real time, finds the patterns and tells your staff what is likely to happen next.
We're looking for federations, clubs and statistical data to support football analysis.
Request a sample reportSee sample work
Founded in 2025 by Mario Casamalhuapa. We work privately with a small number of teams in Mexico and Latin America, never league-wide.
Every number here is measured from the video, as it plays.
Twelve seconds of a recent professional match, read by FF's video model after the match. Blurred and recoloured on purpose: we show the method, not the match.
How we work
From football data services and your club's own numbers, on one match clock.
futbolfutbol AI, on open-weight models we fine-tune ourselves: it connects every source and reads the game.
FF's video model turns the match into coordinates: every player, every quarter of a second.
Before, during and after the game: what happened, what it means and what is likely next.
How it works
Nothing has to change in how your staff works. If it is written down, recorded or exported, we can use it. Every source goes on one match clock, so FF sees minute 63 from every angle at once.
Match and season numbers from football data services, plus your club's own numbers from your statisticians and the organization's.
The file your GPS system already exports after training and games. We read it as it comes, as structured XML or as a table.
The match video, read by FF: every player followed, the key plays found with your staff's rules, drawn and paused.
FF, futbolfutbol's own AI, connects every source, keeps it in order, reads the game as it happens and says what is likely to happen next.
A football person checks every page before your staff sees it.
How they attack, where they concede and the key matchups, ready before the trip.
ExampleShort, timed notes for the analytics team as the game changes.
ExampleThe full story of the game the day after, with the key plays drawn on video.
ExampleWhat your staff gets
Any club can collect or buy the data. The difference is how fast the right number reaches the right person, and what it predicts. These pages come from our sample work, built from public match data only.
How they attack, where they concede, the key matchups and clear points for the staff, ready before the trip.

How the game is turning, minute by minute, with short notes for the analytics team, where competition rules allow.
Every shot, how each goal was made, zones, passes and positions, with the key plays drawn on the video.

Columbus won away with 9 shots against Dallas's 16, scoring from its two biggest chances (xG 0.58 and 0.45).
Open the sample →San Diego had 61.7% of the passes and no shot in the first half, then scored three times between 50' and 64' from 6 shots and 0.60 xG.
Open the sample →
Orlando led 2-0 at half-time from very little, then dropped deep while Chicago moved its attack to one side and came back.
Open the sample →Independent sample work, built only from public match data; not made for, or endorsed by, the clubs shown. Our goal is to work with clubs across the United States and CONCACAF.
Synthetic scenarios
We start from your current roster and your current situation: who is available, each player's GPS load, the opponent's recent form and the schedule ahead. Then we run thousands of synthetic scenarios with established statistical models, so your staff sees the range of outcomes before making the call.
What scorelines are likely?
Goals are rare events, so each team's goal count is modelled with a Poisson distribution around its expected goals for this match. From two numbers we get the probability of every scoreline, and from those the chance to win, draw or lose.
Where does the season end?
We play the rest of the season 10,000 times, drawing each result from that match's probabilities. The spread of final points shows the chance of reaching the playoffs, not a single forecast.
Where should the ball go next?
A possession is a chain of moves between zones. From how often the ball moves on, is lost or ends in a shot in each zone, the model gives every zone a value: the chance the possession ends in a goal from there. The arrow shows the move that adds the most.
How does each goal change the match?
We start from the pre-match estimate of each team's scoring rate and update it minute by minute with what actually happens (a Bayesian update). The line is the chance of winning as the game unfolds.
Which lineup holds up best?
For each lineup option we resample the roster's own past games thousands of times (bootstrap). The dot is the expected goal difference; the bar is the range in 90% of resamples. An option whose range crosses zero is not clearly better.
Synthetic data, computed in your browser: every visit draws a new match. With a club, the same models run on its own roster and data, and an analyst checks every output.
How the models workWhat we build
Nothing has to change in how your people work. We take your data as it is, work on it faster than a staff can by hand, clean it up, and give it back in tools you control. Then we keep improving them, match after match.
We take the files your staff already uses (spreadsheets, apps, GPS exports, notes), check them against each other, fix what does not match and put everything on one match clock. Faster to use, and nothing is lost.
Report rooms, dashboards, scouting lists and match-day screens, in English, Spanish or Portuguese, behind your club's own login. Built around how your staff already works.
FF reads your data on demand: before the match, at half-time and live. It runs on our own private hardware, so results come back fast and your data stays private.
Open-weight models fine-tuned to your club: your game model, your vocabulary, your report style. Every correction your staff makes helps the next report.
Statistics, AI agents and checks work in order: one prepares the data, one builds the forecast, one re-counts every number, one writes, and a football person approves. Every step is logged.
You decide what FF may read, what it may do and who sees the results. Every change is versioned, every forecast is scored after the match, and the tools get better every week.
Our services
Our goals
Science
Established statistical methods, applied the same way every game. FF runs on open-weight models that we hold, run and fine-tune ourselves for football; leading commercial AI models are used only with a club's written agreement.

RPIX · WHITE PAPER · v1.2
How do you compare players competing in different leagues? RPIX combines playing time, production and league strength in a rating from 0 to 100.
Snapshot: 11 November 2026. Agreement with a squad selection is not validation; historical testing is still pending.
PDF v1.2 · 8 pages · English
Partner clubs
We are not building this league-wide. We are choosing a small number of clubs to build a close, long-term relationship with, hand in hand with the analysts, the coaches and the general manager, for a full season.
We work in the background. The result we are after is the club's, on the pitch.
Your club's data belongs to your club. It is never sold or shared with another club, and it goes to a commercial AI service only with your written agreement. The technology and the science are futbolfutbol's.
We bring the match data, FF and the people.
Read the sample agreementQuestions
AI match analysis for professional soccer clubs: opponent previews before the game, live notes during it, and post-game reports and video analysis after it, built from match data, the club's own data, GPS and video on one timeline.
Whatever your staff already has: statistics in any form (an app, a spreadsheet, paper or a voice note), GPS exports as XML or CSV, and match video. We bring the match data.
Yes. It belongs to your club, is never sold or shared with another club, and goes to a commercial AI service only with your written agreement.
FF (futbolfutbol AI), our own. By default, open-weight models that we hold and run ourselves, fine-tuned for soccer. Only with a club's written agreement, leading commercial AI models through their APIs. A football person checks every page before your staff sees it.
Yes. Live analysis is ready to run with a partner club, for the analytics team and, where competition rules allow, the bench.
Per quote. We work to the requirements set by each club, and the price follows the scope you choose.
Request a sample report on one of your recent games. We reply within two working days.
Tell us about your club and we will send a sample report on a recent game.