Temu Product Scraper

Scrape Temu product data reliably. This actor reverse-engineers Temu's anti-content request signer and reproduces it in pure Python, so after a one-time stealth-browser bootstrap it fetches products over plain HTTP — fast and lightweight.

What you get

Two modes:

Mode	What it does
Product feed (products)	Collects Temu's homepage product feed with full data.
Product detail (product_detail)	Enriches a list of product IDs / URLs with server-rendered detail.

Output fields

Product feed rows:

Field	Description
goods_id	Temu product ID
title	Product title
price	Current price (float, in the chosen currency)
market_price	Original / strike-through price
price_str / market_price_str	Formatted price strings
currency	Currency code
discount_text	Discount label (when present)
sales_tip	Sales indicator (e.g. "28 sold")
image	Product image URL
url	Product page URL
source	plp_feed

Product detail rows:

Field	Description
goods_id	Product ID
title	Product title
description	Product description
category	Category breadcrumb (e.g. patio > lawn & garden > outdoor carts)
keywords	Raw category keywords
image	OpenGraph image
url	Canonical product URL
login_or_captcha_walled	true if Temu blocked that page (retried automatically)
source	pdp_ssr

Example input

{
    "mode": "products",
    "maxItems": 100,
    "region": "211",
    "currency": "USD",
    "language": "en",
    "proxyConfiguration": {
        "useApifyProxy": true,
        "apifyProxyGroups": ["RESIDENTIAL"]
    }
}

Product detail:

{
    "mode": "product_detail",
    "goodsIds": [
        "605981362750070",
        "https://www.temu.com/goods.html?goods_id=601099651743804"
    ]
}

Input parameters

• mode — products (feed) or product_detail.
• maxItems — max unique products in feed mode. Temu rotates the feed, so very large runs plateau (~40–160 unique products per session).
• goodsIds — product IDs or URLs for detail mode.
• region / currency / language — locale (defaults: 211 / USD / en).
• headless — run the one-time bootstrap browser headless (default true).
• proxyConfiguration — proxy settings; residential proxies recommended.

How it works

1. A stealth browser loads temu.com once to harvest a live anti-content seed and session cookies.
2. The feed is then fetched via curl_cffi with a pure-Python anti-content minted per request — no browser round-trips.
3. Detail mode reuses the same browser to read Temu's server-rendered product metadata.

Scheduling & webhooks

Run on a schedule and connect webhooks to push new products to your systems as soon as a run finishes. Use the dataset API/exports (JSON, CSV, Excel) to integrate downstream.

Use with AI assistants (MCP)

Prefer to query Temu live from Claude, Cursor or your own agent? Pair this actor with Temu MCP Server, which exposes Temu product search and detail as MCP tools an AI assistant can call directly.

Limitations

• Temu gates keyword search, category browsing and full product detail (price/gallery/reviews) behind login/captcha. This actor delivers the homepage product feed (full data) and SEO-level product detail. It does not log into Temu accounts.
• The feed is a rotating recommendation set, so the number of unique products per session is limited by Temu.

FAQ

Do I need to provide cookies or tokens? No — the actor bootstraps its own session.

Why residential proxies? Temu is aggressively anti-bot; residential proxies make the one-time bootstrap far more reliable.

Are prices accurate? Prices come directly from Temu's feed API and are normalized to the chosen currency.