Chrome extension · Data Cloud

Data Cloud Inspector: a read-only lens on Salesforce Data Cloud

Live October 2026 to now

Chrome MV3 Data Cloud Identity resolution SQL
Type
Chrome extension (Manifest V3)
Stack
Vue, Vite, Tailwind, CodeMirror, dagre
Version
1.0.x
Publisher
Ateko, maintained by me

The problem

Most Data Cloud questions I get in a day are small, but each answer is spread over different Setup pages. By the time four tabs are open, the thread is gone. Users of Marketing Cloud Toolkit who also work in Data Cloud asked for the same thing there: one place to look a person up, read the data model and run SQL.

What it does

  • Profiles: the unified record, its source records, and the match rules beside each record's values. "How they were merged" draws a graph of why each pair matched.
  • SQL runner: tabs, autocomplete, a linter, CSV download and a visual query builder that suggests joins from the org's own relationships.
  • Segments: criteria in plain words, the equivalent SQL, and a rule-by-rule check of why one person is or is not in a segment.
  • Impact: what breaks if you change an object or field, followed up to five levels down to data targets, with a removal plan.
  • Compare: two orgs, or one org over time, with a deployment checklist.
  • Health, freshness and credits: stream freshness, identity merge health and an estimate of the credits a query will consume.

How it is built

Nothing is installed in the org. The service worker uses the Salesforce login you already have in the browser, and the session never reaches the app page. For orgs that block browser sessions, OAuth with PKCE through your own External Client App is the fallback.

"Read-only" is enforced in code, not promised in a README. A guard in the service worker allows only GET requests, plus POST to the Data Cloud query endpoints, and refuses anything else before a request is made. Metadata is cached per org, people's data stays in session storage, and the content security policy only allows calls to Salesforce. Admins can lock the extension down further through Chrome managed policies.

What I learned

  • The Lightning session has no API access. The my.salesforce.com session is the one that works.
  • Query paging has sharp edges. Without a row limit, the first page stops near 48k rows, and queries without ORDER BY page in no stable order.
  • Labels come back HTML-escaped. Apostrophes showed as ' until I decoded them.
  • The query API does not report rows processed. That is why every credit figure in the extension is labelled as an estimate.
  • Segment membership tables use 15-character IDs, while most other places use 18.