Taobao/Tmall Item Detail Data for AI Agents
Give an AI agent a Taobao or Tmall item ID or link and get structured details back: title, price, sales, main images, and SKUs in one shot — no page parsing, no scraper to maintain.

Tool names in this guide are for reference only and may have changed or been renamed. Please use the latest tools in the tool marketplace when implementing.
Key takeaways
- Hand the agent an item ID or link and get structured Taobao/Tmall item details back
- Works by ID or by URL, and can fetch description images separately
- Pulls data through a ready-made endpoint — no page parsing, no anti-scraping to maintain
Tools used
Why pulling Taobao/Tmall item details is such a pain
Getting a Taobao or Tmall item's title, price, sales, main images, and SKUs into clean, structured form looks easy — until you try. Page structures change often, fields hide behind dynamic rendering, and description images are a string of lazy-loaded URLs. Roll your own scraper and you also inherit sessions, pagination, captchas, and anti-scraping — maintenance so costly it breaks every few weeks.
For one or two items, copying by hand is survivable. But the moment you need to batch-ingest, build sourcing comparisons, or feed a downstream analysis, the manual route collapses. The traditional alternative is wiring up the data endpoints yourself — a high bar, and one that shifts every time the endpoint does.
AgentEarth takes a different approach: these capabilities are packaged as ready-made tools. You hand the agent an item ID or a link, describe what you want in one natural-language instruction, and it picks the tools, fetches the data, and organizes it into a structured record.
The workflow in four steps
- Provide the entry point — give the agent an item ID or an item link.
- Fetch structured details — get title, price, sales, main images, and SKUs by ID or by URL.
- Fetch description images — pull the long images from the item description separately.
- Organize and store — have the agent compile a structured record for your database or downstream analysis.
How to do each step
Step 1: Give the agent an ID or a link
Just say what you have. For example: use ae to get this Taobao item's details, ID 1234567890, or use ae to parse the item at https://example.com/item">https://example.com/item." You don't need to know which endpoint runs underneath.
Step 2: Fetch structured details by ID or by URL
- When you have an item ID, use an item-detail tool (such as
ae_tmapi_taobao_item_detail) to retrieve structured details by ID. - When you only have a shared link, use a detail-by-URL tool (such as
ae_tmapi_taobao_item_detail_by_url) to parse the link directly — no need to dig the ID out of the URL yourself.
Both entry points return the same set of structured fields — title, price, sales, main images, SKU variations — so the agent gets clean data rather than a blob of HTML.
Step 3: Fetch the description images
For the long description images embedded in the listing body, use a description-image tool (such as ae_tmapi_taobao_item_desc_images) to pull back the list of image URLs — handy for asset archiving or image-content analysis.
Step 4: Compile a structured record
Have the agent merge the detail fields and description images into one record — written to a spreadsheet or database, for instance. To batch, loop the same chain over a set of IDs or links.
| Approach | Stability | Maintenance | Field completeness |
|---|---|---|---|
| Parse pages / write a scraper | Breaks when the page changes | High — reworked every few weeks | Dig fields out yourself; easy to miss some |
| Use a structured endpoint | Stable; never touches the page | Low — one plain-language call | Title/price/sales/images/SKUs all included |
How to get started
- Sign up for AgentEarth — new users get 1,000 free credits.
- In agents like WorkBuddy or Trae, search for "AgentEarth" and connect in-app; or paste a one-line MCP config into Claude Code / Cursor / VSCode.
- Give instructions in natural language, e.g. use ae to get this Taobao item's details and compile a record.
FAQ
Does it support both ID and link? Both. With an item ID, use an item-detail tool (such as ae_tmapi_taobao_item_detail); with only a shared link, use a detail-by-URL tool (such as ae_tmapi_taobao_item_detail_by_url) so you never have to extract the ID yourself.
Can it fetch description images? Yes. Use a description-image tool (such as ae_tmapi_taobao_item_desc_images) to pull the list of long description image URLs for asset archiving or image analysis.
Which fields come back? Structured details typically include common fields like title, price, sales, main images, and SKU variations. You only need an AgentEarth key and natural-language instructions — no Taobao session to maintain and no API to integrate yourself.
