Your product data is sensitive. We treat it that way. Every layer of Mimir is built to keep your data encrypted, isolated, and under your control.
All content is encrypted with AES-256-GCM before it reaches the database. Field-level — project names, sources, insights, recommendations, chat messages, and business context are each individually encrypted.
Every database query is scoped to your account. No API endpoint, URL, or query path can return another user's data. Enforced at the database layer, not just application logic.
Google OAuth only — no passwords stored. Sessions are database-backed and revocable. Every API route checks authentication before returning data.
Powered by Anthropic's Claude API. Anthropic does not train on API inputs (privacy policy). Your data is never shared with third parties for model training.
Vercel SOC 2 Type II hosting. Neon PostgreSQL with connection-level TLS. Parameterized queries only — no raw SQL. All traffic encrypted in transit via HTTPS with HSTS.
Content Security Policy, X-Frame-Options DENY, strict transport security, content type sniffing prevention, and restrictive referrer policy on every response. Camera, microphone, and geolocation disabled.
Per-user and IP-based rate limiting with tiered limits for AI operations, uploads, and standard requests.
Product analytics track feature usage, not your content. We never log or analyze what you put into Mimir.
Delete any project, source, or conversation at any time. Deletion is permanent — removed from the database, not soft-deleted.
If you have questions about how Mimir handles your data, or if you need specific security documentation for your organization, get in touch. We're happy to discuss our security practices in detail.