Freezetime vs SABR
Freezetime
benginn.github.ioSABR
morat.aiSide by side
| Figure | Freezetime | SABR |
|---|---|---|
| Price from | €4.99/mo | — |
| Free plan | Yes | — |
| Free trial | — | — |
| Open source | Yes | — |
| Web | Yes | Yes |
| Windows | — | Yes |
| macOS | Yes | — |
| Linux | Yes | — |
| API | Yes | — |
| Self-hosted | Yes | — |
| Plans published | 4 | 1 |
| Spec score · Game Replay Analyzers | 9.1 (#2) | 6.9 (#21) |
| Data retention | Archived matches are kept indefinitely, while older seasons receive less weight in tendencies and predictions using a 90-day half-life. | After subscription termination, raw customer data is deleted or de-identified within 60 days of a customer request, subject to stated exceptions. |
| Purpose | Freezetime is a self-hosted analysis platform for Counter-Strike 2 demos. | SABR provides scrim review, opponent preparation, and scouting for professional esports teams. |
| Support | The project directs questions and setup help to GitHub Discussions and bugs or feature requests to GitHub Issues. | Paid subscriptions receive email support with responses within seven business days; the terms make no uptime commitment unless an Order Form says otherwise. |
Plans
In their own words
Freezetime
Freezetime is a free, self-hosted analysis platform for Counter-Strike 2 demos. Its public archive is a read-only site with more than 2,600 professional matches, searchable and grouped by series, and needs no installation or account. For a personal demo, the Analyze page parses one file in the browser using WebAssembly, so the file stays on the user's machine. Processing produces 16 Hz player positions along with kills, grenades, and economy data. Synchronized 2D replays show player health, economy, kill feed, grenade trajectories, smoke, and flash effects. Analysis tools include opponent reports, predictions, ML Lab, leaderboards, match comparison, heatmaps, ghost rounds, Moments, and Pattern Finder. My DB can process private demos without the server keeping a copy. Self-hosting unlocks features marked self-host only, but users must supply their own demos because the repository contains no match data or database dumps. The project says smart features use deterministic methods and locally trained models rather than external AI services. Freezetime is MIT licensed, with no warranty.