
Maximum Data Integrity Instead of Guessing
- Rigorous deduplication:Automatic selection of the most informative “master record.”
- Strict Quarantine: Games with conflicts (color swaps, FIDE ID discrepancies) are not recommended but are isolated.
- Transparent Identities: Phonetic and historical reconciliation with proof of origin in the PGN tags.

Full PGN and SCID support
- Optimized formats: Ready to use as PGN and in the Scid formats
si4andsi5. - Focus on standard chess: Games played exclusively from the starting position (no Chess960 or fragment games).
- Adjusted data: Removal of unnecessary engine evaluations while fully preserving the think times (
%clk,%emt). - Chronological sorting: Sorted deterministically by date, event, location, and round.

Independent
- Focus on Quality: Developed and maintained by Michael Jansen, the sole developer in Germany.
- No data at any cost: Conservative data preparation ensures reliable analyses without incomplete fragments.

Deduplication Logic
- 10-Step processing: Multi-stage pipeline for the precise identification and resolution of duplicate batches.
- Fuzzy Matching & ELO Resolver: Intelligent matching of variant spellings and assignment based on ELO histories and FIDE IDs.
- Subsumption Check: Automatic detection of incomplete games, fragments are replaced with the complete game without any data loss.

Performance & Rust-Backend
- Modern Infrastructure: High-performance data processing with a specially developed Rust pipeline.
- Fast Exports: Efficient storage cleanup and fast loading times despite massive amounts of data.
- Resource-efficient: Minimal memory requirements for smooth operation in Scid and Scid vs. PC.

Monthly release cycles
- Reliable Schedule: New database releases and OTB updates are reliably available in the first week of every month.
- Incremental patches: Monthly update files for directly updating existing Scid collections.
Database version for Scid
- Optimized for Scid and Scid vs. PC (
.si4&.si5). - OTB Database: Over 12 million games.
- Online database: Over 6 million games
Database Version as a PGN File
- Universal format for all common chess GUIs.
- Contains exactly the same cleaned-up dataset.
- Ready for immediate analysis or conversion.
Current Version
Current status:
October 2026
Main Database:
2026-10-05
Incremental Update:
none
Opening Books
ECO/NIC 2025-10-05
License Model (CC BY-NC 4.0)
Lumbra’s Gigabase ist ein freies Projekt für die Schach-Community. Die Nutzung für private Analysen, Vorbereitungen und Training ist unter der Creative Commons Lizenz (CC BY-NC 4.0) vollständig kostenlos. Für die Integration in kommerzielle Schach-Apps oder Datenbanken kontaktiere mich gerne direkt unter moc.esabagigsarbmul@ssenisub
Secure Lichess Integration
- No contamination: To avoid legal disputes, Lichess Broadcast data is strictly excluded from the main database through December 2023.
- License-free: This ensures that the main database remains free of ShareAlike restrictions (CC BY-SA 4.0).
- Separate Download: If needed, the broadcast data is available as a completely standalone, deduplicated download.
Reviews of BuyMeACoffee.com
Note on TWIC
The use of TWIC (The Week in Chess) data was deliberately removed from the Gigabase, as this project relies on donations for its long-term sustainability and should not be undermined by database exports. TWIC is an extremely important project for the chess community.
„Send a £30 donation via Paypal and contact me via email (Email Mark Crowther – moc.tenretnitb@htworcdm) I’ll send you an address for a cbv file of my personal copy of every issue of the games in one database. Over 4 million games. Alternatively subscribe to donate £4 a month“
Partner Showcase
Some links in this Partner Showcase are affiliate links. If you purchase software or subscribe through one of these links, I’ll receive a small commission to support this project, of course, the price remains exactly the same for you.
Thank you for your support!






