Open Science, by Design
Libre Biotech isn't just a research-management tool. It's a statement about how science should be done — openly, reproducibly, and on infrastructure you control. Here's what that means in practice.
The Three Principles
Every design decision in the platform comes back to one of these three commitments. If a feature doesn't serve at least one of them, it doesn't belong here.
Open
The code is AGPL-3.0. The data formats are documented standards, not proprietary blobs. The protocols you publish are public by default, not locked behind paywalls. We build on open shoulders — ISA-Tab, RO-Crate, CWL, the OBO ontologies — and we give back the same way. Open means you can read the source, run your own instance, and take your data with you. Always.
Reproducible
Every result has a trail. Every sample knows where it came from. Every pipeline run records its parameters, its tool versions, and its inputs. The provenance graph isn't an afterthought — it's the backbone. If another lab can't regenerate your figure from your data six months from now, the science didn't happen. We build the infrastructure that makes reproducibility the default path, not the heroic achievement.
FAIR
Findable, Accessible, Interoperable, Reusable. Not a label to slap on a dataset, but a set of engineering commitments: persistent identifiers, ontology-annotated metadata, machine-readable exports, standard workflow descriptions. The platform measures its own FAIR score on every investigation — because what gets measured gets improved, and because AI-ready data is just FAIR data done well.
Our Commitments
Principles are easy to write down. These are the practical commitments that make them real.
Your data stays yours
No terms-of-service clause claiming a licence to your research. No "free tier" that silos your experiments. Full export at any time — ISA-Tab, RO-Crate, PROV-O, plain CSV. If you self-host, you never need to ask us for anything.
No vendor lock-in
Standard schemas (ISA, OBI, EDAM), standard exports, standard pipelines (CWL, Nextflow, Snakemake). Migrating off the platform is a rsync and a SQL dump. Your choice to stay should be based on the tool working well, not on how hard leaving would be.
Credit where it's due
ORCID-linked authorship on every investigation. Per-entity licensing (CC-BY, CC0, or custom). When you publish a protocol or share a dataset, your name travels with it — cited, traceable, attributable. Reputation is infrastructure.
Sovereignty over convenience
Self-host on a spare machine. Run it on your institution's infrastructure. Keep research data inside your jurisdiction's borders. The platform was built to run anywhere a modern PHP stack runs — not to funnel everyone into one cloud.
Community-reviewed protocols
Protocols are versioned, forkable, and open to suggestion. When you improve a method, your fork is visible; when someone else improves your method, you see the diff. Science as collaborative editing, not a one-way broadcast.
Built for AI, honestly
Machine-readable investigation cards. Flat-matrix ML-ready exports. Explicit provenance in every dataset. If AI models are going to consume biology data, the data should arrive with its assumptions, its methods, and its provenance intact — not stripped down to anonymous arrays.