Data Handling

Data handling principles

The DMTR extension follows a stateless architecture in this sprint:

• No local persistence — no IndexedDB, no localStorage for entity data. Every enrichment is fetched fresh from the API.
• No third-party data sharing — all API calls go to the DMTR backend. No data is sent to PitchBook, Google Analytics, or any other service.
• Minimal Chrome storage — only the API URL and session state are stored via chrome.storage.
• No page content capture — the extension reads the DOM to extract entity names but does not store or transmit page content.

Data flow

PitchBook page DOM → Content script (extract name only) → Service worker
    → DMTR API (/v1/intelligence/resolve, /v1/intelligence/enrich)
    → Side panel (render enrichment)

Future: Layer 2 (Personal Knowledge Graph)

When the personal knowledge graph is implemented, the extension will use encrypted IndexedDB for local storage of:

• Entities the user has researched
• Notes and annotations
• Deal pipeline state

This data will be encrypted at rest and never shared without explicit opt-in.