LibreAGI launches private tracker to preserve open-source AI model weights
LibreAGI is a new private BitTorrent tracker where members archive and seed open-weight AI models that are at risk of disappearing. The site rewards seeders with higher credit for rarer files, verifies each model's source commit and license (Apache-2.0, MIT or Qwen Community License), and publishes a signed public catalog of hashes and sources so others could rebuild the swarms independently.
GoKawiil's interpretation of the reporting above, not reported fact.
Open-weight models are sometimes removed by their original hosts or companies, which can cut off access to research and derivative work built on them; a distributed seeding network could make such removals less final. The scarcity-based incentive structure suggests an attempt to direct community effort toward the models most in danger of vanishing rather than ones already widely mirrored. Publishing a signed catalog separately from the tracker itself may be intended to ensure the preservation effort doesn't depend on the site's own survival.
- LibreAGI is a private BitTorrent tracker dedicated to archiving open-source AI model weights.
- Seeding credit scales with scarcity, giving up to 4x rewards for models with only one existing copy.
- Only safetensors/GGUF files under specific open licenses are accepted, verified against source commits, with a public signed catalog for independent recovery.
WD My Book 18TB External HDD — Seeding large model weights long-term means you need serious local storage that won't fail on you. An 18TB external drive gives you ample room to keep copies of safetensors and GGUF files seeding around the clock without worrying about running out of space. It's a practical way to contribute to preserving open models exactly like LibreAGI describes.
See WD My Book 18TB External HDD on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: libreagi.com, 2026-11-12
Published there as: “Libreagi – Save OSS Models”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.