Changelog
12-08-2025β
π New: MCP Server for Coalesce Catalogβ
You can now connect your favorite AI tools directly to Coalesce Catalog using the Model Context Protocol (MCP).
What it does
The MCP server exposes 8 tools that let external AI assistants search and retrieve information from your catalog:
-
π Search tools: Find tables, dashboards, columns, and glossary terms using the same search engine as our AI Assistant
-
π Catalog tools: Access company context, tags, users, and teams
Why it matters
Build your own data agents with your preferred AI tools (Claude, Cursor, Dust, and other MCP-compatible clients). Combine Coalesce Catalog with other MCP servers like dbt, Looker, or Snowflake to create powerful, customized workflows.
β‘ How to get started
-
Generate an API token in your account settings
-
Configure your MCP client with the server URL:
https://api.castordoc.com/mcp/server -
Start querying your catalog through natural conversation
π See the full setup documentation for detailed instructions.
12-04-2025β
π Full Metadata Historyβ
The History tab now logs all metadata changes on assets.
Why you'll care?
-
Traceability for governance: Get a complete audit trail of who changed what and when on tables, dashboards, and knowledge pages.
-
Smoother collaboration: Everyone can quickly verify updates in one place, aligned with what notifications show. No more confusion.

11-28-2025β
πΌοΈ Data Product Images in Catalog Marketplaceβ
You can now add an image to every Data Product in the Catalog Marketplace β a feature many of you have been asking for since the Marketplace launch to make it nicer to scan.
Images can be added by uploading your own or generating one with AI.

π§ͺ In-app Catalog API Playgroundβ
Make it easier and faster for teams to start using the Catalog API directly from the product.
π― What it does:
- New API playground in the app where users can experiment with the Catalog API in a safe, guided way
- Users simply paste the API token provided by the catalog support team and can immediately run real requests
- A pre-configured example query is available out of the boxβclick "Run" to see live results right away
- The query editor supports autocomplete (for example, when typing
foldersand related paths), helping users discover the right endpoints and parameters much faster
π‘ Why it matters:
- Reduces friction for developers and data teams to start integrating with the Catalog API
- Accelerates API-based initiatives by shortening the time from idea to first successful request
π§ Coming soon
- Admin users will be able to generate API tokens directly in the app, then immediately test them in the playground with pre-written queries
- This will provide a fully self-service API onboarding flow: generate token β test requests live β integrate into your own workflows

11-07-2025β
π Search Assistant - Natural Language Filteringβ
Filtering capabilities are now released in the Search assistant!
π― What's new:
Users can now filter search results directly through natural language queries, without needing to switch to manual filters. The assistant automatically understands and applies filters based on:
- OWNER β Filter by business owner (partial or full match)
- TAG β Filter by asset tags
- NAME β Filter by asset name (partial or full match)
π‘ Examples:
- "Which tables are owned by Alice?"
- "Show me dashboards tagged GDPR"
- "Find assets named customer"
This makes the Catalog fully explorable through conversation and significantly speeds up asset discovery! π

10-27-2025β
π Metadata Editor - Filter by Upstream Tableβ
Enhanced the metadata editor with powerful lineage-based filtering capabilities to streamline bulk metadata operations.
π― What it does:
- Added "Ancestors" filter in the metadata editor's table tab
- Filter tables by their ancestors using lineage relationships
- Supports case-insensitive search with flexible naming formats (
database.schema.tableordatabase schema table) - Considers all lineage types:
AUTOMATIC,MANUAL_CUSTOMER,MANUAL_OPS, andOTHER_TECHNOS
π‘ Why it matters:
- Enables data governance managers to efficiently propagate ownership and tags downstream
- Eliminates manual navigation through lineage tabs for each table
- Leverages existing lineage relationships for smarter bulk metadata operations
- Streamlines governance workflows by focusing on dependent assets
ποΈ Metadata Editor - Hierarchical Source/Database/Schema Selectionβ
Improved the metadata editor with intelligent cascading filters for better navigation and selection accuracy.
π― What it does:
- Database options now depend on selected sources
- Schema options now depend on selected databases
- Creates a logical hierarchy: Source β Database β Schema
- Prevents invalid combinations and reduces selection errors
π‘ Why it matters:
- Provides more intuitive and logical navigation through data assets
- Reduces user errors by showing only valid options at each level
- Improves efficiency when working with large, multi-source environments
- Creates a more structured approach to metadata management
10-20-2025β
π Unified Search Experience: ASK + SEARCH in One Placeβ
We've merged AI-powered and keyword-based search into a single, intelligent search bar on the homepage β making it easier than ever to find what you need, however you prefer to search.
