Labsco
MCP SERVER

DataDoe

by Deltologic

Read and act on live Amazon Seller Central, Vendor Central and Ads data without doing the SP-API approval yourself.

E-commerce & Marketplace Operations
Summary
The write path has a dry run and a per-type switch.

Letting a model change live listings and pause campaigns is the risky half of this, and the design takes it seriously: every action type starts disabled, `dryRun=true` validates a payload without executing, and a disabled type still accepts dry runs so you can build the call before you allow it. On the read side, `exports_create` accepting filter groups and aggregations means a complex report is one tool call rather than a chain of them.

What it is

A hosted Amazon server. DataDoe holds the SP-API developer registration, OAuth, refresh tokens and rate limits; you connect your Amazon account to them and point any MCP client at one authenticated URL.

What you get
  • `exports_sources_get` searches pre-built Amazon export templates — orders, sales and traffic, ads performance, inventory, listings, settlements, returns, brand analytics
  • `exports_create` builds a report from any source with SQL-like filters, GROUP BY, aggregations (sum / avg / count / countDistinct / min / max) and DAY / WEEK / MONTH intervals, as CSV or JSON
  • `exports_get`, `exports_raw_url_get`, `exports_raw_download` and `exports_delete` handle the job lifecycle
  • Writing back runs through Actions: `actions_details_schema_get` returns the payload schema, `actions_start` with `dryRun=true` validates, `actions_start` executes, `actions_get` polls, `actions_list` reviews history
  • Actions cover listings (AMAZON_LISTINGS_UPDATE), orders (AMAZON_ORDERS_CANCEL, AMAZON_ORDERS_CONFIRM_SHIPMENT) and Amazon Ads campaigns, ad groups, targets and ads
  • `sellers_and_vendors_list` and `organization_and_subscription_details_get` cover the account; `files_create`, `files_list`, `files_get`, `files_download_url_get` and `files_delete` handle listing and A+ images
  • `datadoe_user_docs_table_of_contents_get` and `datadoe_user_docs_page_get` let the agent look up features on demand
Requirements

A DataDoe account with your Amazon Seller Central or Vendor Central account connected, and an MCP key created in DataDoe's integrations page. The endpoint is https://mcp.datadoe.com/mcp/v1 over streamable HTTP with the key in a `datadoe-mcp-key` header. Nothing is deployed locally. Action types are disabled by default and enabled one at a time in settings.

Setup effort

One command plus a key — npx -y mcp-remote https://mcp.datadoe.com/mcp/v1, then supply credentials