Swantide's Data Dictionary is an AI-generated inventory of every metadata component in your connected Salesforce org. It's the fastest way to see what exists, how each component is configured, and how components relate to each other. The same documentation is what the AI Assistant searches against when answering your questions.
What the Data Dictionary contains
When you connect a Salesforce org, Swantide ingests all metadata and generates documentation for every component, including:
Custom Objects — field counts, descriptions, and relationships
Custom Fields — data types, formulas, picklist values, and where they're used
Flows — trigger conditions, decision logic, and actions
Apex Classes and Triggers — purpose, methods, and dependencies
Validation Rules — criteria and error messages
Page Layouts — field placement and section organization
Permission Sets and Profiles — access configurations
Reports and Dashboards — report types, filters, and groupings
Each component's detail page opens on a Details tab that shows the raw metadata plus a short Swantide-written description in two sections: Business Context and Implementation Details. These descriptions are intentionally a few sentences, not exhaustive — they're summaries for both you and the AI Assistant, not full specs.
Navigating the Data Dictionary
Open the Data Dictionary (in the sidebar under My Catalog; the page is also titled "Documentation").
Use the Filter button to narrow by category group (Automations, Data Model, UI & Experience, Security & Access, Integration, and more). Large orgs can carry tens of thousands of components — filtering keeps things manageable.
Search by Label to find a specific item.
Click into a component to open its detail view, which has three tabs:
Details — the AI-generated description (Business Context + Implementation Details) plus configuration details (fields, criteria, logic)
References — what the component references and what references it
Changes — recent modifications to that component
The Data Dictionary header shows the Last synced timestamp, so you always know how fresh the data is.
Key features
AI-generated descriptions
Every component has an AI-written, plain-English description generated from the actual metadata configuration — not guessed. Expect a few sentences per component, not a full spec. The same documentation powers the AI Assistant, which is why the more components Swantide has documented, the better the AI Assistant gets.
Filtering by category
Use the Filter button to focus on a category group, then browse the component type you care about — for example Flows (all record-triggered, screen, scheduled, and autolaunched flows), Custom Fields across objects, Apex Classes, or Validation Rules.
CSV export
Export the current view with the Download CSV button for offline analysis, team sharing, or migration planning. See How to Export Metadata Documentation as a CSV.
Change Log page
For an org-wide view of what's changed recently, open the dedicated Change Log page (/catalog/changes). You can filter by 24h / 7d / 30d / Custom range, see summary tiles, and use the Ask Swantide button on any change to start an AI Assistant chat pre-loaded with that change's diff context. See How to Use the Change Log to Track Org Changes.
Manual sync
Swantide syncs nightly. When you need an immediate update, there are two manual sync entry points:
Per-component refresh — on any component detail page, click Sync component. Use this when the component is already in Swantide and you just want a fresh pull.
Sync a new component — in the Data Dictionary, click Sync Metadata. The Sync Metadata Component modal asks for a Component Type (e.g.,
ApexClass,StandardValueSet,GlobalValueSet) and a Component API Name. Use this when the component has never been ingested.
See How Often Does Swantide Sync Metadata for the full sync cadence.
When to use Data Dictionary vs. AI Assistant
Use Data Dictionary when… | Use AI Assistant when… |
You want to browse and explore what exists | You have a specific question to answer |
You need a complete, deterministic list of components | You need analysis, recommendations, or comparisons |
You want to export data for offline use | You want to generate documentation, designs, or code |
You need to enumerate (e.g., "list every field on Account") | You want to debug or explain a specific component |
The AI Assistant is agentic — it queries your org's metadata with tools on demand (searching, counting, reading components, tracing references) rather than pre-loading a fixed handful of components. Being specific still gets better answers; it's just not capped at a fixed number of components. It's reliable for counts, bounded object-scoped lists, and tracing what references a single component. For an exhaustive, org-wide list ("find every component that uses X"), the References tab on a component detail page or the Data Dictionary itself is your source of truth — it's built to return every item — while the Assistant gives you a strong, representative answer for analysis. See How to Use the References Table to Understand Metadata Dependencies.
Tips
Start here when onboarding — the Data Dictionary gives you the fastest overview of any org.
Use filters aggressively — large orgs can carry tens of thousands of components.
Check the References tab before you change or delete anything (see How to Use the References Table to Understand Metadata Dependencies).
Match the prompt to the tool. Asking about one specific flow, profile, or field always works. For a guaranteed-complete org-wide inventory, browse the Data Dictionary instead.
Two manual sync entry points — per-component refresh and "sync something new." Don't wait for the next nightly sync if you only need one thing.