Dark Matter Docs

Architecture and capabilities of the DMTR Chrome extension for private market intelligence enrichment.

Overview

The DMTR Chrome extension detects private companies on supported research platforms, resolves them against the Blackmoat intelligence database via the DMTR API, and renders structured enrichment in a side panel. When you visit a PitchBook company page, the extension extracts the entity, calls /v1/intelligence/enrich/:slug, and displays the company profile, transaction history, key people, valuation metrics, and — if DMTR positions exist — transferability status.

The extension is the Layer 1 surface of the contributed data flywheel: free page-level utilities that demonstrate value before asking for anything in return.

Architecture

The extension is a Manifest V3 Chrome extension (packages/extension/) with four components:

Component	File	Role
Background service worker	`src/background/index.ts`	Handles messaging between content scripts and the side panel, manages API calls to the DMTR backend, and controls the side panel lifecycle.
Content scripts	`src/content/index.ts`	Injected on matched sites. Finds the active `SiteAdapter`, calls `extractEntities()`, and sends detected entities to the service worker.
Side panel	`src/panel/`	React app rendered in Chrome's side panel. Three tabs: Know (intelligence), Own (portfolio — placeholder), Signal (market signals — placeholder).
Site adapters	`src/adapters/`	Declarative per-site entity extraction. Each adapter defines URL patterns, CSS selectors for entity extraction, and injection points for UI enrichment.

SiteAdapter interface

Every supported site gets one adapter implementing this interface:

interface SiteAdapter {
  name: string;
  patterns: Array<{ hostname: string; pathname?: string }>;
  extractEntities(doc: Document): ExtractedEntity[];
  injectionPoints(doc: Document): InjectionPoint[];
}

Currently implemented: PitchBook. See PitchBook adapter for details.

Data flow

User navigates to a PitchBook company page
Content script runs pitchbookAdapter.extractEntities(document) → extracts company name
Content script sends ENTITIES_DETECTED message to service worker
Side panel receives entities, calls POST /v1/intelligence/resolve to match the name
Side panel calls GET /v1/intelligence/enrich/:slug for the matched entity
Side panel renders the Know tab with entity profile, metrics, transactions, people, and transferability badge

Three-layer tab architecture

Tab	Layer	Status	Description
Know	Layer 1 — Page utilities	Implemented	Entity enrichment: profile, metrics, transactions, people, transferability
Own	Layer 2 — Personal knowledge graph	Placeholder	Portfolio tracking, deal tracker, relationship graph (requires DMTR account)
Signal	Layer 3 — Collective intelligence	Placeholder	Aggregated market signals, sentiment, deal flow velocity (requires data contribution)

Try it without PitchBook — Start the API with pnpm dev and open http://localhost:3100. The Intelligence Discovery section on the demo page provides the same entity search and enrichment that the extension's Know tab renders. Search for "Saronic" to see the full enrichment panel with transferability badge.

Supported browsers

Chrome and Chromium-based browsers (Edge, Brave). Firefox and Safari support are planned.

What this sprint does not include

IndexedDB / local knowledge graph (Layer 2)
Differential privacy / contributed data (Layer 3)
Bloomberg, AlphaSense, Carta, LinkedIn, or Gmail adapters
MCP server integration
Production authentication (extension uses the sandbox API)

Extension Overview