Philosophy
We enforce "Security by Design" at the database, application, and AI layer.
Report
Request Audit Log01 Multi-Tenancy
Row-Level Security (RLS)
We enforce strict logical separation via Postgres RLS policies. Every database query—including Vector Search—automatically filters by your org_id at the engine level, making cross-tenant data access mathematically impossible.
Tech Stack
- + Supabase (Postgres 15)
- + pgvector (Embeddings)
- + Vercel Edge Functions
02 AI Guardrails
Grounding
All agent outputs are anchored in "Today's Date" and your specific GTM DNA context to prevent hallucinations.
Permissions
Agents operate on a "Least Privilege" basis. They cannot spend budget or send emails without an explicit scoped token.
Isolation
LLM Context Windows are ephemeral. We do not use your data to fine-tune shared foundation models (e.g., Gemini 2.0 Flash).
03 Access Control
Authentication
We use Supabase Auth for secure identity management, supporting Magic Links and Social Login (Google). All sessions are JWT-encrypted.
Encryption
Data is encrypted at rest (AES-256) and in transit (TLS 1.3). API Keys for third-party tools are stored in a separate encrypted vault.