Skip to main content
The Catalog Search Tool allows DinoAI to search across your full data catalog, helping you discover the assets you need without leaving the Code IDE. This tool lets DinoAI find dbt models, sources, tests, and macros, along with assets from third-party integrations, so you can quickly locate the right resource before building or extending a model.

​
Capabilities

The Catalog Search Tool exposes a single underlying operation: search_catalog searches the data catalog across all dbt™ assets (models, sources, seeds, snapshots, tests, macros, and exposures) as well as connected integrations such as Looker, Tableau, and Fivetran. It takes a text query and a result limit (default 5, at most 25). Each result carries the asset’s name, type, key, description, tags, and schema, plus the columns that matched when the match was on a column. Long descriptions are shortened and matched columns are capped at 10, so an answer stays small. For a dbt™ asset the key is its unique_id, which is what the dbt Discovery tools and the Column Level Lineage Tool take.
Scope is your production catalog: the dbt™ project of your default Bolt environment, built from its latest manifest.json and catalog.json, plus every connected integration, whose data refreshes daily.

​
Where you can use it

  • Copilot, in the right panel of the Code IDE.
  • Agents: programmable agents, the Slack agent, and Bolt agents. In the agent builder the tool sits in the Catalog & lineage group; in an agent YAML, name it search_catalog in tools.list.
  • MCP clients, over the MCP server.

​
Using the Catalog Search Tool

  1. Open DinoAI, or message an agent that has the tool
  2. Describe the asset you want to find
  3. Add your prompt describing what you want DinoAI to do with the catalog results
  4. Review DinoAI’s findings and apply them to your development work

​
Example Use Cases

Searching the Data Catalog Prompt
Result: DinoAI queries the catalog across dbt models, sources, and third-party integrations, returning matching assets with their descriptions and tags. Finding where to start, from Slack Prompt
Result: the agent searches the catalog for the dashboard and its models first, instead of searching the repository file by file, then follows their lineage from the unique_ids it found. Checking for Existing Assets Before Building Prompt
Result: DinoAI searches the catalog for matching dbt models, sources, and BI assets so you can reuse existing work instead of duplicating logic.

​
Working with Other Tools

The Catalog Search Tool works well alongside DinoAI’s other capabilities to support your full development workflow:
  • Combine with the Column Level Lineage Tool to trace the upstream or downstream dependencies of an asset you’ve found in the catalog
  • Combine with the SQL Execution Tool to validate the structure of a model or source surfaced by a catalog search
  • Combine with the Google Docs Tool or Notion Tool to cross-reference catalog assets against your data modeling specs
  • Use alongside Git Lite to commit and push catalog-informed changes before opening a pull request

​
Best Practices

Use catalog search to explore before you build. Run a catalog search before creating new models to check whether a similar asset already exists in your workspace. Use descriptive search terms. Search by business concept (e.g., “marketing attribution”, “customer churn”) to surface relevant assets across both dbt and third-party tools. Account for refresh cadence. Catalog data reflects the latest production manifest.json and catalog.json; third-party integration data refreshes daily, so very recent changes may not yet appear.