Full Disclosure
I'm Federico, founder of RaceTagger. I'll be transparent about where RaceTagger has advantages and where TrackAction.AI might be a better fit for your specific situation. Both tools exist because manual race photo tagging is a painful, time-consuming process that deserves a better solution.
Quick Overview
| RaceTagger | TrackAction.AI | |
|---|---|---|
| Type | Desktop app | Desktop app + cloud AI |
| AI Processing | Cloud-based (Gemini AI) | Local detection + cloud number reading |
| Pricing | Free tier + one-time packs from €39 | $10-299/month subscription or one-time packs |
| Free tier | 500 tokens + 100 free/month | 100 credits on signup |
| Sports | Motorsport, running, cycling, triathlon | Motorsport focused |
| Platforms | Windows, macOS, Linux | Windows, macOS |
| CSV matching | ✅ Built-in | ✅ Separate CSV module |
| RAW support | ✅ All major formats | ✅ Via embedded previews |
How Each Tool Works
RaceTagger's Approach
RaceTagger sends your photos to Google Gemini AI for analysis. The AI examines each image, detects race numbers, and returns structured results. RaceTagger then matches detected numbers against your CSV participant list and writes the complete metadata (number, driver name, team, category) to your files as IPTC/EXIF keywords or XMP sidecars.
Key technical features:
- Smart Matching with multi-evidence scoring (number, sponsor text, team names)
- Temporal clustering propagates identifications across burst sequences
- OCR correction with 100+ confusion patterns (6↔8, 1↔7)
- Configurable sport categories with optimized AI prompts per discipline
- Processes NEF, CR2, CR3, ARW, and all major RAW formats
TrackAction.AI's Approach
TrackAction.AI uses a hybrid local+cloud system. A local AI model first detects subjects in your images (people vs vehicles) without uploading anything. For vehicle photos, it crops just the relevant section and sends that small crop to cloud AI for number reading. People photos are tagged locally as "portrait" at zero credit cost.
Key technical features:
- Local-first detection (privacy advantage: full photos never leave your computer)
- Crop-and-send approach for race number reading
- Portrait auto-tagging distinguishes people from vehicles locally
- XMP pre-filter lets you process only starred/flagged photos to save credits
- Multi-core processing for parallel analysis
Feature Comparison
| Feature | RaceTagger | TrackAction.AI |
|---|---|---|
| Race number detection | ✅ Gemini AI | ✅ Cloud AI on crops |
| CSV participant matching | ✅ Built-in, automatic | ✅ Separate module |
| Burst sequence propagation | ✅ Temporal clustering | ❌ Per-image analysis |
| OCR confusion correction | ✅ 100+ patterns | Not documented |
| Multi-evidence matching | ✅ Number + sponsor + team | Number-based |
| Portrait detection | ❌ | ✅ Local, free |
| Pre-filter by star rating | ❌ | ✅ XMP pre-filter |
| IPTC/EXIF writing | ✅ Direct to file | Via XMP |
| XMP sidecar writing | ✅ | ✅ |
| Offline mode | ✅ (syncs when online) | Partial (local detection only) |
| Multi-sport support | ✅ Motorsport, running, cycling, triathlon | Primarily motorsport |
| Linux support | ✅ | ❌ |
| Log visualizer/dashboard | ✅ | ❌ Not documented |
Pricing Comparison
This is where the two tools differ significantly.
RaceTagger — One-Time Token Packs
- Free tier: 500 tokens on signup + 100 free analyses/month
- Starter Pack: €39 for 5,000 tokens
- Professional Pack: €49 for 10,000 tokens
- Studio Pack: €99 for 25,000 tokens
- Tokens never expire. No monthly commitment.
TrackAction.AI — Monthly Subscriptions + One-Time Packs
- Free: 100 credits on signup
- Starter: $10/month for 1,000 credits
- Pro: $39/month for 5,000 credits
- Business: $69/month for 10,000 credits
- Agency: $299/month for 50,000 credits
- One-time packs also available ($9-89)
- Credits rollover up to 12 months with active subscription
Cost Per Photo Analysis
| Volume | RaceTagger Cost | TrackAction.AI Cost |
|---|---|---|
| 1,000 photos | €7.80 (Starter) | $15 one-time or $10/mo |
| 5,000 photos | €39 (Starter pack) | $49 one-time or $39/mo |
| 10,000 photos | €49 (Professional) | $89 one-time or $69/mo |
| 25,000 photos | €99 (Studio) | ~$222 in one-time packs |
| 50,000 photos/year | ~€198 | $299/mo ($3,588/year) or ~$445 one-time |
Key difference: RaceTagger's one-time packs with no expiration are significantly cheaper for seasonal photographers who don't need monthly subscriptions. TrackAction's monthly plans suit photographers who process consistently every month and want automatic credit allocation.
Privacy & Data Handling
RaceTagger: Photos are sent to Google Gemini AI for analysis. The full image (or embedded preview for RAW files) is transmitted to the cloud for processing. Results are returned and stored locally.
TrackAction.AI: Full photos never leave your computer. Only small cropped sections of vehicles are sent to cloud AI. Local AI handles the initial subject detection. This is a meaningful privacy advantage if you're concerned about uploading client photos to cloud services.
Verdict on privacy: TrackAction.AI's crop-and-send approach offers better privacy by design. If your clients or contracts require that full images stay local, TrackAction has the edge here.
Accuracy
Both tools report similar accuracy ranges:
- RaceTagger: 85-95% on clearly visible numbers (98% achieved at Ferrari Finali Mondiali Mugello 2025)
- TrackAction.AI: 80-90% first-pass accuracy (per their documentation)
Both recommend a manual review pass after AI processing. Real-world accuracy depends heavily on image quality, number visibility, lighting, and angle.
RaceTagger's temporal clustering can improve effective accuracy by propagating identifications across burst sequences — if one frame in a 10-frame burst has a clear number, all 10 get tagged. TrackAction processes each image independently.
Who Should Choose Which?
Choose RaceTagger if:
- You're a seasonal photographer who doesn't want monthly subscriptions
- You shoot multiple sports (motorsport + running + cycling)
- You process large batches (10,000+ photos per season) and want the lowest cost per photo
- You want temporal clustering for burst sequence handling
- You need Linux support
- You value multi-evidence matching (sponsor text, team names beyond just numbers)
- You want to try extensively before paying (500 free tokens + 100/month)
Choose TrackAction.AI if:
- Photo privacy is your top priority (only crops leave your computer)
- You want portrait auto-detection as part of the workflow
- You prefer monthly subscription with predictable costs
- You primarily shoot motorsport only (not multi-sport)
- You want to pre-filter by star rating before processing to conserve credits
- You process photos every month consistently
Consider both if:
- You cover many events and want to compare results on your specific images
- Both offer free tiers — test each on the same event folder and compare
FAQ
Can I use both tools on the same photos?
Yes. Each tool writes to different metadata fields by default. However, running both on the same batch doubles your cost and time. Better to pick one and stick with it.
Do both tools work with Lightroom?
Yes. Both write XMP metadata that Lightroom Classic reads on import. RaceTagger also writes directly to IPTC/EXIF fields.
Which has better customer support?
Both are relatively small teams. RaceTagger offers direct support via email and a community of beta testers. TrackAction.AI has documentation and email support. Neither has 24/7 phone support — these are specialized niche tools.
What if I shoot both motorsport and running events?
RaceTagger supports multiple sport categories with optimized AI prompts for each discipline. TrackAction is focused primarily on motorsport. For multi-sport photographers, RaceTagger offers more flexibility.
Try both free and decide for yourself. RaceTagger: Download free → — 500 tokens on signup + 100 free analyses every month. TrackAction: visit trackaction.ai for their 100-credit trial.