Security
Sovereign by architecture, not policy.
Most 'private AI' vendors promise not to look. Locaible can't look — your data and models live on your hardware. We built the system so trust is structural.
Local execution
All inference runs through Ollama on your machine. There is no Locaible inference endpoint, ever.
Zero document upload
Your RAG documents are indexed locally by the installer. They never touch our infrastructure.
Approve-on-demand support
Remote access goes through a Tailscale tunnel that only opens when you press Approve in your dashboard. Sessions are logged on your machine.
Metadata only
Our backend stores account, plan, and agent metadata (name, type). No conversations, no embeddings, no documents.
RGPD & EU AI Act ready
Data residency is your machine. Right-to-erasure is a delete button. We provide DPA and risk-classification documentation on request.
Open and auditable
The installer scripts and agent runtime are inspectable. No magic, no hidden network calls.