LineSquire
Open navigation menu
Product design commitments

Safety and privacy

Honest identity. Bounded authority. Visible data choices.

LineSquire is designed to say what it is, do only what the subscriber permits, and leave a clear record of every consequential decision. This page describes intended controls, not a legal certification.

Required first turn

“Hi, I’m an AI assistant calling for the subscriber about the approved purpose.”

AI identity disclosed

Consent event recorded

Tools remain allowlisted

Privacy defaults

Collect less. Explain more.

Call data is sensitive personal data. The intended defaults keep raw audio out of storage and make retention a visible user choice.

Raw recording off

Recording and transcript controls are separate, visible settings.

AI disclosed first

The assistant identifies itself before collecting the purpose of a call.

Permission before commitment

Fees, terms, sensitive facts, and out-of-plan choices pause for approval.

Export and deletion

Call-level deletion, account export, and deletion jobs are part of the intended product.

Hard boundaries

Some actions stay out of reach.

User approval does not turn a prohibited action into an allowed one. Launch scope intentionally excludes high-risk and abusive calling.

Bulk, sequential, predictive, or campaign calling

Sales, political persuasion, fundraising, or debt collection

Voice cloning, impersonation, spoofing, or deceptive identity

Passwords, full payment-card data, bank credentials, or one-time security codes

Autonomous purchases, contracts, rights waivers, or financial transfers

Emergency response, medical decisions, or professional legal or financial advice

Emergencies are not delegated.

LineSquire is not an emergency service. The intended flow tells the caller to contact local emergency services, may offer a permitted transfer, and sends a high-priority alert without diagnosing or promising a response.

Caller instructions cannot widen permission.

A caller cannot reveal private subscriber data by claiming to be the subscriber, change the approved plan, expose hidden instructions, or add tools. Verification and tool permissions stay outside the voice model.

This is product design guidance, not legal advice. Exact call categories, disclosures, consent handling, retention, and launch jurisdictions require qualified legal review before outbound service.

Always AI. Approval before action. Delete your data anytime.

Control should be visible before it is needed.

Use the demo shell to preview disclosure, approval, and receipt states without placing a real third-party call.