Privacy Policy
Last updated: 14 October 2026
1. Who We Are
Memnar is a product of Mission Relearn (KvK: 99384949), based in the Netherlands. Mission Relearn is the data controller for the personal data described in this policy. For any privacy question, or to exercise your rights, email annet@memnar.ai.
2. What We Collect
- Account data: your email address and display name, when you sign up.
- What you upload: the documents (PDF, DOCX, TXT, MD) and voice recordings you provide so Memnar can build your AI-self.
- Your generated AI-self: the set of markdown files Memnar writes from your uploads. These can include passages quoted word for word from your own material.
- Work patterns (only if you connect a tool): if you connect Memnar to an AI coding tool, we distil patterns from those sessions into your files. We store the distilled patterns, not the raw session transcripts.
- Billing data: if and when payments are enabled, a payment identifier and billing details. Payments are currently switched off, so we collect no billing data today.
- Connection data: a hashed token if you connect Memnar to another tool.
3. How We Use Your Data, and Our Legal Basis
- To build your AI-self from what you upload — on the basis of your consent, which we ask for explicitly before your first upload.
- To provide and run the service (your account, storage, export and deletion) — on the basis of our contract with you.
- To send transactional email (sign-in links, account notices) — on the basis of our contract with you.
- To keep the product secure and working — on the basis of our legitimate interest in a reliable, safe service.
Your uploads may contain special-category personal data (for example health information in a document you wrote). We process this only with your explicit consent. Please do not upload other people’s confidential or special-category data without their permission.
4. How Your Documents and Voice Are Processed
Your uploaded documents are read once to build your AI-self, then deleted from our servers. We do not keep copies of your uploads. Voice recordings are sent to Mistral for transcription and Memnar does not store them afterwards. Mistral keeps the data it receives through its API, including recordings and uploaded text, for up to 30 days for abuse monitoring, then deletes it. Your generated AI-self stays in your account until you delete it. Every quote in your files is checked against your own uploads before it is written.
5. Sub-processors
We use a small set of processors, all operating under data-processing agreements. Since our 2 July 2026 move to European vendors, every processor that handles your content processes it in EU or EEA regions. Some of them run on infrastructure of US-owned cloud providers in those regions; see section 6.
| Processor | What it does | Where | Data |
|---|---|---|---|
| Supabase | Account, database, storage, authentication | EU (Ireland) | All categories |
| Mistral AI | Reads your text to build and refine your AI-self, and transcribes your voice recordings | EU/EEA regions (French company) | Uploaded text, voice recordings, distilled patterns. Mistral keeps API data for up to 30 days for abuse monitoring, then deletes it |
| Resend | Sends sign-in and account emails | EU (Ireland) | Email address |
| Cloudflare | Hosts the site and serves static assets | EU network | None directly (your data lives in Supabase) |
| PostHog | Product analytics, and in the app also session replay and error reports, which PostHog analyses automatically to suggest improvements | EU (Germany, Frankfurt); the automatic analysis uses AI providers (OpenAI, Google, Anthropic, Microsoft) in the United States, under the EU-US Data Privacy Framework and standard contractual clauses | Usage events and bare metadata; in the app tied to a pseudonymous account ID, never your email, uploads or AI-self content. Session replays are fully masked: every piece of text and every input is hidden before the recording leaves your browser, so a replay shows where you clicked, never what you wrote |
| Stripe | Billing, only once payments are enabled | EU / US | Email and billing details (not in use today) |
| Substack | Newsletter delivery, only if you tick the optional newsletter box at signup (off by default); unsubscribe via the link in any issue | US | Email address (opt-in only) |
Our AI processors do not use your content to train their models. We do not sell your data, and we do not share your personal data with third parties for marketing.
6. International Transfers
Your content is processed in EU or EEA regions: Supabase in Ireland; Mistral, a French company, with infrastructure in EU/EEA regions, partly with US cloud providers; Resend in Ireland. We removed the previous US processors at the 2 July 2026 cutover. Stripe (which can involve US processing) is present in our stack but payments are switched off, so no personal data flows to it today; we will confirm EU terms before enabling payments.
Two transfers do exist. If you opt in to the newsletter, Substack (US) receives your email address. And PostHog stores usage data in Germany, and since 4 September 2026 it also analyses that data automatically to suggest product improvements. That analysis runs on AI providers in the United States: OpenAI, Google, Anthropic and Microsoft (Microsoft also processes in Germany). PostHog participates in the EU-US Data Privacy Framework, and standard contractual clauses apply on top of that. PostHog contractually forbids these providers from using our data to train or fine-tune their models. Nothing you upload and nothing from your AI-self is in that material: see section 9 for exactly what it contains.
7. Your Rights (GDPR)
Under the General Data Protection Regulation, you have the right to:
- Access your personal data
- Correct inaccurate data
- Delete your data (“right to be forgotten”)
- Export your data in a portable format
- Object to processing
- Withdraw consent at any time
You can export or delete everything yourself, anytime, in Settings. To exercise any other right, email annet@memnar.ai. We respond within one month.
8. Data Retention
- Uploaded documents: deleted by Memnar after your AI-self is generated; text sent to Mistral is kept there for up to 30 days, then deleted.
- Voice recordings: not stored by Memnar after transcription; Mistral keeps them for up to 30 days for abuse monitoring, then deletes them.
- AI-self and work patterns: kept until you delete them or close your account.
- Account data: kept while your account is active.
- Payment records: once payments are enabled, retained for 7 years as required by Dutch tax law.
- Backups: held on our providers’ normal cycle (roughly 7 to 30 days). Deleted accounts are not restored from backup.
9. Cookies
On this site we use functional cookies only, for sign-in and session management. We do not use tracking or advertising cookies. Our analytics (PostHog, hosted in the EU) runs cookieless here: nothing is stored on your device, and we never send it your email address or any content you upload.
Inside the app (app.memnar.ai) PostHog also records masked session replays and error reports, so we can see where the product breaks or confuses people. Masked means every piece of text and every input field is hidden in your browser before the recording is sent, and we never record the contents of network requests. A replay shows the shape of the page and where you clicked; it does not show what you wrote, uploaded, or what your AI-self says. The screens that show your AI-self, your patterns and your uploaded file names are left out of recordings entirely. PostHog analyses this material automatically to propose product fixes. That analysis is about the product, not about you: it makes no decisions about any individual person.
In the app this runs on our legitimate interest in improving a service you rely on, which means you can object at any time and we stop. Switch Usage analytics off in your app Settings and nothing is recorded at all, or email us and we do it for you. You never have to give a reason. Replays and error reports are kept 30 days, then deleted.
10. Automated Decision-Making
Memnar does not make automated decisions that produce legal or similarly significant effects about you. Building your AI-self is a tool you direct, not a profiling decision made about you.
11. Supervisory Authority
If you believe we handle your data incorrectly, you can file a complaint with the Dutch Data Protection Authority (Autoriteit Persoonsgegevens) at autoriteitpersoonsgegevens.nl.