Cookies and data

Last updated May 27, 2026

We keep our cookie use small on purpose. Most of what's listed below is functional, meaning it remembers a choice you made (theme, sidebar state, what you typed into the home hub). Our analytics provider is cookie-less. We don't sell data and we don't share it with advertisers.

You can change your choice on this page at any time. If you'd rather skip the granular controls, you can

or
.

Your preferences

Essential

Required for Meridian to work at all: sign-in, security, and remembering your consent choice. Can't be turned off.

Always on

Functional

Remembers what you typed and how you laid out the app: theme, sidebar state, home-hub form values. Doesn't track you.

Analytics

Anonymous usage measurement via Plausible (cookie-less, IP-masked, EU-hosted). Helps us see which pages are useful without identifying anyone.

AI personalization

Lets Meridian remember your prior conversations and vault context to personalise future AI assistance across cases. Off by default. Core AI features (chat, readiness checks) work either way.

What's actually set

Every cookie we set, what it does, and how long it lasts.

Essential

NamePurpose
_meridian_sessionKeeps you signed in across pages.
cookie_consent_v1Stores your choice on this page so we don't ask again.

Functional

NamePurpose
home_destination / home_purpose / home_residence / home_arrival_date / home_return_dateRemembers what you typed on the home hub so you don't re-enter it.
activity_rail_state / sidebar_stateRemembers whether you collapsed the sidebar or activity rail.
localStorage.themeRemembers light or dark mode if you switched from system default.

Analytics

NamePurpose
(none)We use Plausible, which is cookie-less and IP-masked. No cookies are set on your device for analytics.

AI personalization

NamePurpose
(none)No cookies set for this category. The toggle controls whether your past conversations and vault are used to personalise future AI responses across cases.

About AI

Meridian is built on AI. We use third-party language models from Anthropic and OpenAI to help check documents, suggest readiness improvements, and draft letters. We do not train AI models on your data. AI outputs are checks and suggestions. You and the consulate make the actual decisions.

Personal data is removed before content is sent to AI providers (passport numbers, dates of birth, financial identifiers). Whether your prior conversations and vault are used to personalise future AI responses is controlled by the AI personalization toggle above. Core AI features work either way.

Contact

Privacy questions go to data.privacy@mail.usemeridian.app. Security reports go to data.security@mail.usemeridian.app.