AI productivity and autonomous schedule intelligence engineered for students and professionals. Combines natural-language instruction reasoning with pure mathematical interval algebra to eliminate LLM hallucinations, guarantee zero calendar overlaps, and protect user privacy.
Engineered from first principles with formal safety constraints, truthfulness, and privacy by design.
Pure Python interval algebra calculates mathematically conflict-free slots around fixed lectures, meetings, and buffer windows. Eliminates LLM time hallucinations.
Consequential external actions (such as sending emails or publishing events) require explicit user confirmation. Actions are staged, verified, and audited.
Google Gemini intelligence with an automated deterministic rule engine fallback. Never invents missing deadlines or fake task durations.
Least-privilege OAuth 2.0 integration with automatic token encryption at rest using AES-128 Fernet and automated token refresh.
Cryptographically verified JWT sessions enforced by PostgreSQL Row Level Security (RLS) policies. User data is strictly isolated.
Native Python MCP SDK server exposing typed tools for AI agent integrations (Antigravity IDE, Claude, Cursor) with idempotency protection.