list_amc_templates
The AMC queries SellerSheet runs for every store whose AMC is on, and the warehouse table each one fills.
readSellerSheet
Read this before writing your own AMC SQL: the answer may already be synced — query its table with query_report_data (get_table_schema gives the columns) instead of running ads_create_workflow_execution.
Call it
In an AI client (MCP)
{
"name": "list_amc_templates",
"arguments": {}
}From code (Tool API)
curl -X POST https://api.sellersheetai.com/api/v1/tools/list_amc_templates \
-H "Authorization: Bearer $SELLERSHEET_API_KEY" \
-H "Content-Type: application/json" \
-d '{}'Early access — the Tool API is open on request, contact us.
Arguments
| Name | Type | Description |
|---|---|---|
key | string | one template, e.g. 'AMC_NTB_BY_CAMPAIGN'; omit for all of them. |
Result
Every tool answers {notification, data, human_action}; data carries:
data.templates[]— {key, version, question, how_to_read, sql, data_sources, mode, cadence, period_days, synced_table, dimensions, metrics, hidden_column, enabled_by_default, notes}.data.templates[].synced_table— the rpt_amc_* table; one row per store, country, period_start and group of dimensions.data.templates[].how_to_read— units and caveats — read it before quoting a number.
Gotchas
- Each period has ONE row with hidden=true: everything AMC suppressed under its privacy floor, summed. Totals include it; per-group answers leave it out. period_start is the first marketplace-local day of the period (7 or 28 days, period_days), not a day of activity. mode scheduled = Amazon runs it weekly; on_demand = every 4 weeks, 15 days after the period ends (attribution settles). A template with enabled_by_default false is not synced unless turned on.