Build useful AI without giving up control.

A practical project, born in France

loqy began with a simple frustration: capable AI should help with real work without requiring people to hand over every document, decision, or workflow to a remote platform.

Its roots are Breton, its daily life is Parisian, and its ambition is broader: make powerful work software understandable, accountable, and usable on your own terms.

Privacy is an architecture choice

Local and private by default is not a badge or a settings toggle. It shapes where work runs, which data can leave the machine, how actions are authorized, and which evidence remains attached.

Sovereignty here means practical choice: choose the model, see what is sent, keep a local path, inspect the work, and retain the ability to move.

Open source by design

The public code makes the product’s claims inspectable and gives the project a durable base beyond one vendor or one person. Source, releases, and technical documentation are available in the public repository.