Privacy Policy
Cairn is built by a solo developer who believes in straightforward data practices. This policy explains what data Cairn collects, why, where it goes, and how you control it.
The Short Version
- Your engineering models live on your device. They reach Cairn's servers only for a project you back up to your account: a project started in the interview backs up as you go (always, when you start it on a phone), and you can switch backup on or off for any project.
- When you use AI features, your prompts and model context are sent to third-party AI providers for processing.
- Your account email and usage metrics are stored in our database for authentication and billing.
- We don't sell your data. We don't run ads.
What Cairn Collects
Account Information
When you create an account, Cairn stores your email address (for authentication via magic link sign-in), a display name (derived from your email, editable by you), and your account tier and usage metrics. This data is stored in Supabase, hosted in the United States.
Project Metadata
When you create a project, Cairn registers a lightweight record: project name, node count, and timestamps. Your actual model data stays in your browser's local storage (IndexedDB) unless you back the project up.
Model Data (Backup)
Backup is set per project. A project you start in the interview backs up to your account as you go: the box is ticked by default on a computer, where you can untick it, and backup is always on for an interview started on a phone, because that is how you finish it on a computer. While backup is on, an unfinished interview is saved to your account a few seconds after each answer. For any other project, backup is off until you turn it on in Settings.
While backup is on, Cairn uploads project snapshots (JSON), binary assets, and ChangeSets to Supabase. A project you never back up never leaves your browser.
The Description You Start With
If you create your account from "What are you building?", the description you typed is stored with your account until your interview starts, so it can open on whichever device you finish signing in on. It is cleared once the interview has it.
AI Feature Usage
When you use AI features, context from your model is sent to Anthropic (Claude) for processing. For image generation, text prompts are sent to Google (Gemini). Cairn logs token counts and estimated costs to meter the early-access allowance.
BYOK users: Your prompts go directly from your browser to the AI provider. They never pass through Cairn's servers. Your API keys stay in your browser's local storage.
The simulation notebook always uses your own key, with Anthropic, OpenAI or xAI, whichever you choose, and sends the project data you allow straight from your browser to that provider.
Analytics
Cairn uses PostHog for product analytics across the site and the app: page views, clicks, the steps of using Cairn (such as starting an interview or applying a change), and session recordings in which every input and all text are masked. When you are signed in, these are linked to your account and its email. Events describe what you did (kinds, counts and ids), not what your model says: no model data or project content is sent to analytics. Vercel Web Analytics counts page views. PostHog honours Do Not Track.
Where Your Data Lives
| Data | Location | Access |
|---|---|---|
| Account email, profile | Supabase (US) | You, Cairn admin |
| Model data (local) | Your browser | Only you |
| Model data (cloud sync) | Supabase (US) | You, Cairn admin |
| AI prompts (managed) | Anthropic/Google | Transient processing |
| AI prompts (BYOK) | Direct to provider | You and provider |
| Notebook prompts | Direct to Anthropic, OpenAI or xAI (your key) | You and provider |
Data Ownership
You own your models. Your system models, requirements, interfaces, designs, and all project content belong to you. You can export everything at any time via the built-in .cairn.zip export.
Data Retention and Deletion
Contact greg@getcairn.dev to request account deletion. All associated data (profile, projects, cloud data, usage logs) will be removed. Local browser data is under your control.
Third-Party Services
| Service | Purpose |
|---|---|
| Supabase | Authentication, database, storage |
| Anthropic | AI model processing (Claude) |
| AI image generation (Gemini) | |
| Vercel | Hosting, Edge Functions, Web Analytics (page views) |
| OpenAI, xAI | The simulation notebook, on your own key |
| Resend | Transactional email |
| Mailtrap | Delivering feedback you send from the app |
| PostHog | Product analytics |
Security
All data in transit is encrypted via TLS. Database access is protected by row-level security. Server-side API keys are stored as environment variables. BYOK keys never leave your browser. If you discover a vulnerability, contact greg@getcairn.dev.
Contact
For privacy questions, data requests, or account deletion: greg@getcairn.dev
This privacy policy is written to be read by humans, not lawyers. If something is unclear, ask.