Where does Footlab AI's data come from?
Footlab AI relies on two sources and two only: API-Football, a commercial football data provider used under a paid licence, for matches, line-ups, statistics, standings, squads, transfers and injuries; and Wikidata, a knowledge base released under the free CC0 licence, for historical honours predating the provider's coverage. No data is extracted from a third-party website without permission.
In total, the database holds 1,561,628 matches and 493,978 players across 171 countries, with honours going back to 1,900.
What exactly are the sources?
Two sources feed the database, each with an explicit legal framework. A third route, web search, never feeds the database: it is used only to verify a probable line-up or a piece of news at the moment a question is asked.
| Source | What it provides | Framework |
|---|---|---|
| API-Football | Matches and scores, line-ups, match and player statistics, standings, squads, transfers, injuries, and published 1X2 odds, which feed the probability calculation and are never displayed. | Paid commercial licence (Ultra plan). Redistribution of the raw database is forbidden: Footlab AI publishes answers and aggregates, never a dump of the database. |
| Wikidata | Honours and competition winners going far beyond the provider's statistical coverage, plus entity identifiers linking a player or club to its public record. | Public domain (CC0). Free reuse, attribution nonetheless stated here. |
| Web search (at question time) | Probable line-up announced by the press, news and rumours on paid plans. | One-off consultation of public pages to verify a fact. Nothing is stored in the database, nothing is republished verbatim. |
Footlab AI performs no automated extraction of third-party websites to build its database. No data is taken from a competitor, a statistics site or a betting operator.
Which competitions are covered, and since when?
The table below shows the real coverage of the main competitions, read straight from the database when this page is rendered. The “detail” column gives the share of matches for which line-ups and match statistics are available, not just the score.
| Competition | Since | Matches | Detail |
|---|---|---|---|
| Serie AItaly | 2,010 | 6,461 | 95 % |
| La LigaSpain | 2,010 | 6,460 | 95 % |
| Premier LeagueEngland | 2,010 | 6,460 | 95 % |
| Serie ABrazil | 2,010 | 6,410 | 99 % |
| Major League SoccerUSA | 2,012 | 6,343 | 98 % |
| Ligue 1France | 2,010 | 6,176 | 94 % |
| EredivisieNetherlands | 2,010 | 5,334 | 94 % |
| Süper LigTurkey | 2,011 | 5,254 | 95 % |
| BundesligaGermany | 2,010 | 5,218 | 95 % |
| Primeira LigaPortugal | 2,010 | 4,950 | 95 % |
| UEFA Europa LeagueWorld | 2,014 | 4,583 | 97 % |
| UEFA Champions LeagueWorld | 2,011 | 3,551 | 96 % |
| Liga MXMexico | 2,016 | 3,536 | 96 % |
| UEFA Europa Conference LeagueWorld | 2,021 | 2,439 | 96 % |
| Euro ChampionshipWorld | 2,008 | 477 | 100 % |
| World CupWorld | 2,010 | 360 | 100 % |
The database holds far more competitions than those listed here: this table is a representative sample, not an inventory.
How often is the data refreshed?
Each family of data has its own frequency, chosen according to how fast it actually goes stale.
| Data | Frequency |
|---|---|
| Scores and events of live matches | Every minute |
| League standings | Hourly |
| Injuries and unavailabilities | Daily |
| Elo ratings | Daily |
| Squads | Every two days, and up to ten days after the transfer window closes |
| Transfers | Daily, with web verification during the transfer window |
| Prediction accuracy | Continuously, no earlier than three hours after each kick-off |
What are the known gaps?
This section exists because a complete football database does not exist. Here is what Footlab AI does not know, written down plainly rather than discovered halfway through an answer.
- Detailed statistics before 2008: missing. Goals, assists and minutes per player do not go back beyond the provider's coverage. Honours, on the other hand, go much further back thanks to Wikidata — a 1955 title is known, the 1955 team sheet is not.
- Provider expected goals (xG): filled in on 2.1% of the matches in the database, and no longer updated since 20 June 2026. The prediction model therefore uses official xG where it exists and estimates it from shots on target everywhere else; no Footlab AI page displays an xG block, for lack of coverage.
- Referee of an upcoming match: known before kick-off for only about 3% of fixtures, often the day before. When it is, it only feeds the probability of the number of cards.
- Official line-ups: published roughly one hour before kick-off. Before that, Footlab AI works from a probable line-up, flagged as such.
- Lower divisions and youth competitions: matches and scores are present, individual statistics far less often than in major leagues.
- A figure may differ from another source depending on the counting method used (play-off matches, own goals, stoppage-time minutes). Footlab AI states the source of every figure, so you know where it comes from.
How can a wrong data point be reported?
A data error can be reported to support@footlab.ai, stating the match, player or club concerned. Corrections coming from the provider are picked up at the next refresh; those coming from Wikidata are fixed at the source, in Wikidata, which benefits everyone.
Footlab AI (footlab.ai) is an independent software product. It is unrelated to FOOTLAB World, the indoor football centre franchise, and unrelated to Footy Labs.