Coach's phone
Shot from behind the baseline. Far-side bounces are hard to see, so placement is tagged in zones you agree, not guessed to the centimetre.
Tennis · Match tagging service
Every point tagged. Every pattern visible.
Send match video. Get every serve, return, rally and outcome back as data, tagged to your definitions and checked by a reviewer, ready for coaching or your app.
COACHES · ACADEMIES · TENNIS APPS · AI TEAMS
POINT 47 · TAGGING
POINT LOGRALLY 5 · WON · WINNER
- 1SERVE1st · T · deuce court
- 2RETURNBackhand · deep cross
- 3SERVE +1Forehand · inside-out
- 4RETURN +1Forehand · short
- 5WINNERForehand · down the line
Illustrative point. Real tags follow your definitions.
66%
of points at the 2024 Australian Open ended within four shots.
That was the men's tournament average, as reported by the ATP Tour. Serve, return and the next two shots decide most points, so that is where detailed tagging pays off first. Your own players' matches may differ, which is exactly why they are worth tagging.
What we tag
Four layers. Pay for the ones you need.
Each layer answers different questions and costs different effort. Pick a layer to see what it adds to a single point.
BARS = RELATIVE TAGGING EFFORT PER MATCH
LAYERS 1–2 ON ONE POINT
Illustrative point on a regulation court. Each layer builds on the ones below it.
FOR · 01
Coaches
Point-by-point stats on your players without spending your evenings tagging.
FOR · 02
Academies
The same definitions applied to every player, every week, so progress is comparable.
FOR · 03
Tennis apps
Consistent tagged matches to power stats features or check your automated output.
FOR · 04
AI teams
Frame-accurate ball events and shot labels to train and evaluate tennis models.
Sample output
Data a coach can use tomorrow.
One row per point for stats and spreadsheets, one record per shot for apps and models, with video timestamps so every row links back to the moment it describes.
match.csv
| pt | serve | dir | result | rally | won | ending |
| 45 | 1st | wide | in | 7 | B | unforced error · BH |
| 46 | 2nd | body | in | 3 | A | winner · FH |
| 47 | 1st | T | in | 5 | A | winner · FH |
| 48 | 1st | T | ace | 1 | A | ace |
| 49 | 2nd | body | in | 9 | B | forced error · FH |
Example structure with synthetic points and figures, not a real match.
Example CSVCan you trust hand-tagged data?
Yes, with clear definitions.
When a 2010 study compared coaches' video coding of serve landing positions with Hawk-Eye, the correlation was r = 0.998 (Hizan, Whipp and Reid). Harder calls, like forced versus unforced errors, depend on how well the definition is written. So we write them down before anyone tags.
Read our guide to tagging tennis matchesLAYER 1
Primary annotator
Tags every point in the agreed layers, at the frame each event happens.
LAYER 2
Specialist reviewer
Checks every tag against the video and the written definitions; failed tags go back.
LAYER 3
Senior auditor
Samples the match for consistency across sets and signs off before delivery.
How it works
Match in. Data out.
- 01
Send match video
From a phone on the fence, a fixed court camera or a broadcast. Full matches or selected sets.
- 02
Agree layers and rules
Which of the four layers you need, and written definitions for the hard calls, like forced versus unforced errors.
- 03
Tag and review
A tennis-literate tagger logs every point, and a reviewer checks each tag against the video before delivery.
- 04
Receive data
One row per point or per shot, as CSV or JSON, ready for a spreadsheet, your app or a model.
Footage we handle
Any camera. Any court.
Academy courts
Many players and courts in one recording, so each match is identified and tagged as its own file.
Broadcast
Replays and cut-aways interrupt points, so replays are flagged and never counted as new shots.
Fixed club camera
Wide lenses shrink the ball, so ball events are tagged on the frame of contact with the racket or court.
Pricing
Priced per match. Tested for free.
PER MATCH
Set by the layers you choose
Point structure and outcomes cost far less than placement on every shot. We quote after a short call about what you want to answer.
How our pricing worksFREE PILOT
One of your matches
One clip, up to about 10 minutes, tagged in the layers you pick and returned in 3 business days. No contract, no payment.
Questions
What coaches ask us.
What should we tag first?
Start with point structure and outcomes, plus placement on the first four shots of each point. That covers the shots that decide most points without paying for every rally ball.
How do you decide forced versus unforced errors?
With a written definition agreed before tagging starts. We can use yours, adapt an open standard such as the Match Charting Project codes, or write one with you.
Can you tag video filmed on a phone?
Yes, as long as the ball is visible for most of the point. Anything out of view is marked as such rather than guessed.
Can the data be frame-accurate for model training?
Yes. Ball events are tagged at the exact frame of contact or bounce, which is the level of detail tennis ball-tracking and event models need.
How is tennis tagging priced?
Per match, by the layers you need and your volume. We quote after a short call. New clients start with a free pilot.
Our footage includes junior players. Is that a problem?
No, but we agree how the footage and data are handled in writing before any work starts, and sign an NDA if you need one.
Free pilot
Stop tagging at midnight.
Send us a match. You coach; we turn the video into numbers you can act on.