Privacy Policy

1. Data Ingestion & Code Parsing

RepoMind indexes public GitHub repositories. When you submit a URL, the system clones files transiently to parse logic boundaries. Raw code is chunked and vectorized; we do not store full code documents permanently.

2. Vectors & Semantic Memory

Vector embeddings representing your codebase structures are stored in a secure instance of a Qdrant vector store. These are tagged and associated with your workspace session token and are deleted periodically to protect space.

3. Account Metadata

For registered accounts, we store only your email address and login timestamps to verify usage bounds and session scopes. For guest users, a transient day token checks query limits. We do not sell or monetize personal metadata.

4. Third-Party Connections

Generative completions and semantic routing query services are processed using Google Gemini neural models. We follow zero-data retention APIs where model providers do not train on code context blocks sent through queries.

Return to Portal