Skip to main content
TruckerFetch exports pin your selection to a published source release. Use an export for a CSV workflow and keep the saved request key so retries refer to the same logical job.

Request a catalog export

Use exports:write and contacts:read to create the export. Generate an idempotency key once, save it with the request body, then submit:
A successful request returns 202 with the job in data. Keep its ID. The edition scope requires edition_id and current access to that date. The list scope requires an owned list_id. The fixed dashboard sample is not an external API export.

Monitor and download

With exports:read, GET /exports/{id} to read job state. Wait until it is ready, then GET /exports/{id}/download to read its manifest. Download also requires contacts:read. The manifest’s data contains the export id, row_count and ordered parts. Each part has an index, length, volume and sha256. Retrieve each part with the same bearer credential at /exports/{id}/parts/{part} and verify its length and SHA-256. Keep binary files out of an assistant’s context. Access is rechecked for downloads. An expired artifact returns 410; a changed suppression or projection can require rebuilding. Never treat an old download or job receipt as permanent authorization.

MCP owner review and retries

Hosted MCP request_export can return an owner approval link. Open that link while signed in, review the selection, then retry the same tool arguments and idempotency_key. A plain success message from an assistant does not replace the product’s owner review. For REST or MCP timeouts, reuse the same key and selection. Inspect the saved result or existing export before starting a new request. See errors for conflict and recovery behavior.