LibreLLM is built on the conviction that AI must be developed in the open, auditable, replicable, and governed by the community.
This is our manifesto:
🧠 Open Model
We release full model weights, training scripts, and architectures. No hidden layers. No obfuscation. No black boxes.
📚 Open Dataset
We curate public, auditable datasets. They are compliant with privacy regulations such as GDPR, and filtered to exclude toxic or copyrighted content.
🔧 Open Tools
We provide scrapers, tokenizers, converters, and preprocessing pipelines. Anyone can reproduce, extend, or audit our workflows.
🚀 Open Pipeline
From raw data to deployed models, we share the complete process including training, evaluation, and inference. Nothing is held back.
🔐 Open Governance
LibreLLM is community-driven. Decisions are transparent, results are reproducible, and the roadmap is open to all contributors.
← Back to homepage