datawebotA product-data loop for your agent
Answer a product question from a bounded extraction.
Updated · datawebot team
Keep the right fields.
Workflow field map| Field / step | Input | Rule |
|---|---|---|
| Task identity | run.id + idempotency key | Save both before yielding. |
| Result pages | products + next_cursor | Continue until the cursor is null. |
| Answer | variants + provenance | Cite the observation; preserve unknown values. |
Which variants does this example mark available?
Small / Navy, Medium / Navy are marked available in this synthetic example. 1 example value is unknown.
No live source, price observation or inventory state is represented.
Before you use the output
The saved example reports availability, not inventory quantity. Reordering, schedules and price changes require your own systems and authorization.
Check one record first.
Start with the smallest useful scope. Compare the returned record with its source, confirm variant identity and inspect missing fields before applying the mapping to a larger dataset.
Keep extraction warnings and the retrieval timestamp with downstream output. When a required field is missing, leave it unresolved or return to the source; do not fill it from an assumption.
Agent setup prompt
set up https://mcp.datawebot.com/setup.md
What this does: your coding agent reads the setup page, asks for your email, waits while you approve the sign-in email from datawebot, saves its key to a private file and connects to the MCP server at https://mcp.datawebot.com/mcp. You never paste a key into the chat.
Then ask, for example: "Find Sony WH-1000XM5 prices in the US and list the sellers with links."
The agent uses search_products to find products across online shops, then get_product_offers for each seller's price and a direct link to the shop.
Results are best-effort snapshots with a retrieval time. They are not verified or live prices.