Rastro MCP

List all catalogs for the authenticated organization.

Get a single catalog by ID, including metadata and item counts.

Delete a catalog (irreversible). Requires explicit confirmation phrase.

Get catalog schema definition with field types, descriptions, scopes, and workflow metadata.

Get catalog taxonomy with hierarchy, attributes, and inheritance.

Set the catalog's quality prompt (used by the judge tool and readiness checks). Pass the full prompt text to replace the current one.

Set the catalog's markdown context (catalog_md). This text is auto-injected into enrichment and mapping prompts. Pass the full markdown to replace.

Get the catalog's markdown context (catalog_md).

Query raw catalog rows with pagination, text search, sorting, and product/variant awareness.

Bulk upsert catalog items. Each item dict should include __catalog_item_id (database UUID) for updates, plus the fields to set/update. Returns summary with items_processed/created/updated/failed counts.

Get a single catalog item by ID with full data, entity type, and taxonomy attributes.

Update a single catalog item's data directly. Use for small, targeted edits (1-5 items). For bulk changes, use the snapshot-diff-stage pipeline instead.

List activities for a catalog with optional status/type filters.

Get a single activity by ID with full metadata.

Get staged changes for a pending activity. Returns paginated list of before/after data pairs.

Build a local HTML artifact for visually inspecting a catalog, an activity's staged changes, or arbitrary rows, then optionally open it in the browser. Pass 'rows' for quick previews without a catalog.

List catalog snapshots for rollback/history.

Create a manual snapshot for rollback safety.

Restore a catalog to a specific snapshot.

Run regex content checks across all items in a catalog. Pass `rules=[{name, pattern, fields, mode?, case_insensitive?}]` for custom checks, or `use_preset='sunco_corona'` for a ready-made ruleset (no Corona, no CL-codes, no MR-dash, no 'in' as a unit

Duplicate a catalog schema and optionally copy source items.

Save an activity as a reusable workflow template (csv_importer -> custom_code).

Create a custom transform activity with staged changes, script provenance, and audit metadata. Validates the bundle, stages all changes into a single pending-review activity (chunked internally if needed), and opens the dashboard review URL.

Map source items to a target catalog schema using AI enrichment. Forces web_search=false.

Judge catalog rows for data quality. Supports multimodal: pass explicit image URLs via 'images' to evaluate whether images match row data. Uses the catalog's schema and quality_prompt automatically when catalog_id is provided.

Download an external image and host it on Supabase Storage. Returns the hosted URL.

Submit an image editing job. Supports: generate, edit, inpaint, bg_remove, relight, upscale.

Get status/progress/result of an image editing run.

List image editing runs with optional filters.

Open a local browser UI to review image generation results. Pass run IDs from service_image_run and optional context per run (product title, finish, etc.). Supports reviewing many runs in parallel.

Export catalog rows + schema to local files (parquet/csv) for Python transforms. Uses fast parallel pagination with retries and safe fallback.

One-command staging: compute diff from before/after datasets and create one pending-review activity. For product_grouped catalogs, validates that every variant's product_id has a matching product parent row.

Compute row/field diff between before/after local datasets. Outputs staged_changes.jsonl for activity creation.

Validate a transform bundle before activity creation. Checks files, schema compatibility, row counts, and policy rules.