Security

Your prompts never cross our servers.

Privacy at Laivara is not a settings page or a compliance checkbox. It is a set of architectural choices made before any feature was built.

The extension captures session metadata — which AI tool, for how long, how often. Never what you typed. Never the output. Never the context.

How your data moves

Your AI session

ChatGPT · Claude · Gemini · Perplexity

What you do

Laivara extension

Captures: tool, duration, timestamp, session count

What we see

Your account

Patterns, weekly insights, controls, exports

What you get

Prompt content

What you type — never captured, never transmitted

×

Metadata flows up

Prompts never leave device

How it is built

Four principles built into the architecture.

01By default

Metadata-first

The product model tracks usage patterns, not private prompt content. Your ideas stay yours by design, not by policy.

02Always on

Device-aware sessions

Per-device authentication, session visibility, revocation, and refresh handling. You always know what has access.

03User-controlled

Consent controls

Clear scopes, permissions, exports, deletion requests, and retention controls. Nothing is buried or locked behind support.

04Architectural

Team boundaries

Admin views are bounded and designed to avoid surveillance. Team visibility has hard limits built into the architecture itself.

Our commitment

What we will never do.

These are not aspirational. They are constraints baked into how the product is built.

×

Read, store, or process the content of your prompts.

×

Sell or share behavioral data with third-party advertisers.

×

Silently change what data is collected without notice.

×

Use your data to train AI models without explicit consent.

×

Hide privacy controls behind paid plan walls.