Skip to main content
The only category write surface exposed to MCP agents. Category writes are TSV-based, round-tripped with export_categories — the UI and REST API support finer-grained edits, but MCP intentionally keeps the surface narrow. This tool is Write scope.

import_categories

Import category definitions from TSV text. Existing slugs are updated (display_name, icon, color, sort_order, hidden). New slugs are created. Missing slugs are not deleted. Parents must appear before children in the input. The typical workflow is round-trip: call export_categories to get the current state, edit the TSV, then call import_categories with the edited content.
Merging categories. Set the merge_into column on a source row to a target slug to merge the source into the target. All transactions and rule mappings from the source are reassigned to the target, then the source is deleted. Useful for simplifying a complex taxonomy without losing transaction categorization.

Parameters

content
string
required
TSV content. Columns: slug, display_name, parent_slug, icon, color, sort_order, hidden, merge_into.

Example input

Author the TSV first — the table below shows what each row looks like before encoding (tabs are real tab characters):
Then encode the TSV into the content field on the tool call. Tab characters become \t and newlines become \n:

Example output

TSV format reminder

  • Tab-separated. Quotes are not required.
  • Empty cells for optional fields are allowed — the server keeps the current value.
  • sort_order is an integer. hidden is true / false.
  • When merge_into is set on a row, all other fields on that row are ignored — only the source slug and the target slug matter.
Last modified on June 25, 2026