1. Get your footage ready
You can analyze a match from three sources: a video file you upload, a direct video URL, or a YouTube link. Uploads can be up to 500 MB in MP4, AVI, MOV, MKV or WebM format. For YouTube videos, PitchView lists the available download qualities and pre-selects the recommended one; you can pick a higher quality for better detail or a lower one for speed.
The ideal recording is a single wide-angle camera that captures most of the pitch. Mount it high if you can β in a stand, on scaffolding or on a tall tripod β so play can be followed across the field rather than just around the ball. Keep the camera steady and avoid zooming in and out during play.
Full matches are fine; there's no need to cut the footage into halves first. Broadcast footage works too: the football analysis is instructed to ignore replays, so a goal shown three times is logged once.
2. Choose the sport and the teams
Pick the sport from the list, or β for a YouTube link β press Auto-detect and PitchView works it out from the video's title and details, using AI when the keywords aren't conclusive. Before the analysis starts you confirm the sport, because it decides what the AI looks for: football footage is analyzed against a football-specific list of events and rules, not a generic one.
Next, pick the two teams. Teams live in your Teams list, where each one has a name, short name, country, colours and a player roster. Choosing them means every event is attributed to the right side by name, and it's optional if you just want a quick look.
3. Choose standard or parallel analysis
Standard analysis processes the whole video in one pass. It's ideal for highlights, short clips or a single half of a small-sided game.
Parallel analysis splits the match into short, slightly overlapping segments β 45 seconds each by default β and analyzes several at once (three by default). Events start appearing while the rest of the match is still being processed, a live counter shows how far along it is, and you can cancel at any point. For a full 90-minute game it's much faster than a single pass. If you're not sure, use parallel analysis for anything longer than about 20 minutes.
If the match is being broadcast live, there's a third option, stream analysis, which follows the game as it happens. It's covered in our guide to standard, parallel and live analysis.
4. What the AI looks for
For football, PitchView logs eight types of event. Goals come with the scorer, the assist provider and the type of finish β header, volley, tap-in or a strike from outside the box β plus how the move was built: counter-attack, possession play or set piece. Shots are recorded on and off target, including efforts that hit the woodwork.
Set pieces cover corners, direct and indirect free kicks, and penalties awarded. Discipline covers yellow cards, red cards (straight or second yellow) and significant warnings. Defensive actions capture key blocks, goal-line clearances and last-man tackles.
Offsides are logged when a flag stops play. VAR is followed from the initial check to the on-field review and the final decision, overturned or upheld. Substitutions record who came off and who came on, whether the change was tactical or forced by injury.
The AI is instructed to leave out anything it isn't sure about rather than guess, and to skip throw-ins unless they lead directly to a goal. You can see exactly what it was asked to do: the prompt used for the analysis can be previewed, and the raw AI response is available after the run.
5. Read and filter the timeline
When the analysis finishes, the events appear next to the video with a timestamp, the team, the event type and a short technical description β for example, "Counter-attack goal, low drive to the bottom-left corner by #10". Click any event and the video jumps straight to it.
A full match can produce dozens of events, so the list can be searched and filtered by event type, by team, by review status, and by whether an event came from the AI or was added by you.
6. Review, correct and add your own
AI output is a starting point, and PitchView is built for reviewing it. Every event carries a review status β Pending, Correct, Wrong, Uncertain, Duplicate or Partial β which you set as you work through the list, with an optional comment, so you always know what's been checked. You can edit an event's time, type, team or description, delete the ones you don't need, and add manual events for anything you want to track that the AI doesn't cover.
When you need a different focus β pressing triggers, one player's involvement, how your team restarts after conceding β write your own custom prompt and re-run the analysis with it. You can also choose the AI model in your settings: Flash is faster and cheaper, Pro gives higher-quality results.
7. Export and come back later
Results can be exported β as an Excel spreadsheet from standard analysis, or as JSON from parallel analysis β for your own reports, clip lists or data tools. Every analysis is also saved to your history, with the video details and the model settings that were used, so you can reopen it, keep reviewing, or re-analyze it later.
How coaches use the results
For post-match review, filter to goals and shots and walk through each chance with the players involved. For set-piece work, pull every corner and free kick, for and against. A run of last-man tackles in one area of the pitch usually shows where the back line is being exposed. For opponent scouting, running the same analysis on an opponent's recent matches gives a quick picture of where they create chances and how they concede.
What AI analysis does and doesn't do
PitchView detects and timestamps match events. It doesn't track player positions, measure possession or generate heatmaps. What it removes is the slowest part of video analysis β finding the moments β so the expertise you bring goes into understanding them.