Jumia Scraper

Scrape Jumia across 9 African markets — Nigeria, Egypt, Kenya, Morocco, Ghana, Uganda, Côte d'Ivoire, Tunisia and Algeria — into structured product rows with price, old price, discount, brand, seller, official-store flag, rating and availability, plus a separate customer-reviews dataset.

Built for African e-commerce price monitoring, marketplace seller research, competitive assortment analysis and catalogue ingestion. HTTP-only: no headless browser, no captcha solver, and robots-compliant by design.

What you get

Monitor prices across 9 markets at once

country, countryName, currency, price, oldPrice, discountPercent and priceFormatted on every row. Currency is per-market (NGN, EGP, KES, MAD, GHS, UGX, XOF, TND, DZ), so you can compare local pricing without FX conversion — and flag currency alongside price when you do aggregate.

Separate official stores from marketplace sellers

isOfficialStore, sellerId and sellerName distinguish brand-authorised sellers from third-party resellers, which is the split that matters for price-quality analysis. brand lets you compare a brand's own store against its resellers.

Get verified-purchase review data

includeReviews: true fills the Reviews dataset with rating, title, text, author, date and isVerifiedPurchase — and isVerifiedPurchase means you can restrict sentiment analysis to confirmed buyers instead of counting drive-by ratings.

Ingest clean multi-market catalogue data

sku (Jumia's config SKU, the stable product key), productId, productUrl, images[], categoryPath[], availability and sourceUrl form a ready-to-load product table.

Know exactly how a row was obtained

sourceType (listing vs product) is on every row. Listing rows come from search/category/seller cards; product rows come from full product pages and additionally carry description and specifications. sourceUrl tells you which page produced the row.

How it works

Jumia's listing pages embed their product payload in the server-rendered HTML, so the Actor fetches pages over plain HTTP with CheerioCrawler — no browser is launched. One listing page yields about 40 products, which is what maxPagesPerQuery is calibrated against.

This Actor treats Jumia's robots.txt as a real constraint rather than an obstacle to route around, and three inputs exist specifically to keep it compliant:

• categoryPaths is preferred over searchQueries. Jumia's robots.txt disallows the ?q= search URLs but leaves category paths open. searchQueries still works, but categoryPaths is the compliant choice — the Actor's own source comments say so.
• botContact is baked into the User-Agent. Jumia's robots.txt permits scraping only for a bot that identifies itself with a reachable owner, and blocks fake user agents. Leaving botContact empty produces a run that identifies as generic — set it to your URL or e-mail.
• maxRequestsPerMinute is capped at 200. That is the ceiling in Jumia's robots.txt. Requests are issued serially from one IP address — this is a deliberate design decision, not a performance oversight.

The result is a slower run than a parallel crawler, and an unambiguously allowed one. If you need throughput, raise maxRequestsPerMinute toward (never above) 200, and accept the trade-off explicitly.

Input

Field	Type	Default	Notes
markets	string[]	["ng"]	Market codes: ng, eg, ke, ma, gh, ug, ci, tn, dz.
searchQueries	string[]	[]	Free-text search terms. Note: ?q= search URLs are disallowed in Jumia's robots.txt — prefer categoryPaths.
categoryPaths	string[]	[]	Category paths such as /phones-tablets/. Preferred and robots-compliant.
botContact	string	""	Your URL or e-mail, baked into the User-Agent so the bot identifies itself with a reachable owner.
maxRequestsPerMinute	integer	120	Requests per minute, ceiling 200 per Jumia's robots.txt.
startUrls	array	[]	Plain Jumia URLs, or { url } objects from the request-list editor.
maxItems	integer	100	Hard cap on products in the Products dataset.
maxPagesPerQuery	integer	1	Listing pages per source. 1 page ≈ 40 products.
includeReviews	boolean	true	Collect reviews into the Reviews dataset.
maxReviewsPerProduct	integer	20	Reviews per product (10 per reviews page).
proxy	object	{ useApifyProxy: false }	Standard Apify proxy picker. See the compliance note below before enabling.

markets, searchQueries, categoryPaths and startUrls are additive — use any combination, but supply at least one. countries remains available as a legacy alias for markets.

Output

Products dataset — 25 fields, one row per product:

country, countryName, currency, title, sku, productId, productUrl, sourceUrl, price, priceFormatted, oldPrice, discountPercent, brand, sellerId, sellerName, isOfficialStore, rating, ratingCount, availability, images, categoryPath, sourceType, description, specifications, scrapedAt

Reviews dataset — 13 fields, one row per customer review:

country, countryName, sku, productTitle, productUrl, rating, title, text, author, date, isVerifiedPurchase, sourceUrl, scrapedAt

Dataset views ship as Products (default) and Reviews.

Field notes that change how you read the data:

• price/oldPrice are numbers; priceFormatted is the rendered local string. discountPercent is null when there is no markdown, not 0 — the opposite of the Hepsiburada Actor, so don't reuse that logic here.
• ratingCount is the count of verified ratings, not total reviews. The Reviews dataset can hold more rows than ratingCount suggests.
• description and specifications are only present on sourceType: "product" rows. On listing rows they are null. To get specs you must fetch product pages.
• country is the market code (ng), countryName is the readable name. Carry both — country is the join key.
• sku is Jumia's config SKU and the stable key across markets; productId is the numeric id from the URL.
• date in reviews is Jumia's own DD-MM-YYYY format, not ISO. Parse it deliberately.

Example input

Compliance-first category walk:

{
  "markets": ["ng", "ke", "eg"],
  "categoryPaths": ["/phones-tablets/", "/laptops-computers-accessories/"],
  "botContact": "https://example.com/bot",
  "maxRequestsPerMinute": 120,
  "maxItems": 300,
  "maxPagesPerQuery": 3
}

All 9 markets, reviews off, price monitoring only:

{
  "markets": ["ng", "eg", "ke", "ma", "gh", "ug", "ci", "tn", "dz"],
  "categoryPaths": ["/phones-tablets/"],
  "includeReviews": false,
  "maxItems": 500,
  "maxPagesPerQuery": 2
}

Seller and assortment research:

{
  "startUrls": ["https://www.jumia.com.ng/product-category/phones-tablets/"],
  "includeReviews": true,
  "maxReviewsPerProduct": 20,
  "maxItems": 150
}

Example output

{
  "country": "ng",
  "countryName": "Nigeria",
  "currency": "NGN",
  "title": "Samsung Galaxy A54 5G (128GB) - Dual SIM",
  "sku": "SI842G7AFDBEA5YEOVB",
  "productId": "428719203",
  "productUrl": "https://www.jumia.com.ng/samsung-galaxy-a54-5g-128gb-dual-sim-428719203.html",
  "sourceUrl": "https://www.jumia.com.ng/phones-tablets/",
  "price": 289000.0,
  "priceFormatted": "₦289,000",
  "oldPrice": 335000.0,
  "discountPercent": 13.73,
  "brand": "Samsung",
  "sellerId": "seller-jumia-ng-1042",
  "sellerName": "Jumia Store NG",
  "isOfficialStore": true,
  "rating": 4.6,
  "ratingCount": 412,
  "availability": "In Stock",
  "images": ["https://storage-jumia.jumia.com/w/280x280/pri/428719203.jpg"],
  "categoryPath": ["Phones & Tablets", "Mobile Phones", "Smart Phones"],
  "sourceType": "listing",
  "description": null,
  "specifications": null,
  "scrapedAt": "2026-10-04T11:48:03.916Z"
}

Integrations & automation

Call it over the Apify API, from Python/JavaScript, the CLI or MCP. Datasets export as JSON, CSV, Excel or XML.

For price monitoring, a daily run over the same categoryPaths per market gives a stable comparable set — 1 listing page is ~40 products, so maxItems: 200 and maxPagesPerQuery: 5 is a comfortable daily cadence. For assortment analysis, run weekly with includeReviews: false to cut cost roughly in half, then filter on isOfficialStore to separate brand pricing from marketplace pricing.

Typical pipeline: run on schedule → upsert on sku + country → compute price and discountPercent deltas → alert on target-price hits → roll up per brand and sellerName. Because requests are serial by design, run markets as separate concurrent runs rather than one wide run if you need to move faster — that is the compliant way to gain parallelism.

Related Actors

Actor	Why
scrapers_lat/jumia-products-scraper	The established Jumia Actor in the Store, 160 users, 5.2K runs. The reference point here.
ahmed_jasarevic/hepsiburada-scraper	Structurally identical problem in a different market: Turkish marketplace with multi-seller, official-store and delivery fields.
ahmed_jasarevic/notino-scraper	Retail catalogue extraction with brand, price and discount modelling, for the Italian market.
ahmed_jasarevic/amazon-bestsellers-scraper	Same retail-data job at global scale — useful for schema and price-tracking patterns.
ahmed_jasarevic/ikea-scraper	Single-retailer product and stock extraction, another comparable reference implementation.

FAQ

How does the price compare to other Jumia scrapers?

scrapers_lat's Jumia Actor lists from $6.80/1K products, while this one is $0.90/1K — about 7.5× cheaper. Reviews are also billed at $0.0009 each here. Prices are a snapshot and providers change them; check the live table on each Actor page before committing.

Is it really allowed to scrape Jumia?

More so than most. Jumia's robots.txt permits scraping only for a bot that identifies itself with a reachable owner and blocks fake user agents — so this Actor exposes botContact precisely so you can put your own URL or e-mail in the User-Agent, and it runs serially from a single IP at or below the 200 requests/minute ceiling the file specifies. It also warns you that ?q= search URLs are disallowed while category paths remain open, which is why categoryPaths is the recommended input.

To be precise about what that does and does not mean: a compliant User-Agent and rate limit make automated access tolerated under Jumia's stated policy. They do not grant permission to resell the data, and they are not a substitute for a commercial agreement. Jumia's Terms of Service still govern your use.

Why is categoryPaths recommended over searchQueries?

Because Jumia's robots.txt disallows the ?q= search URLs but leaves category paths open. searchQueries still works if you need it, but categoryPaths is the compliant way to get the same product data.

Why is it slower than other scrapers — no parallelism?

Deliberate. Requests are issued serially from one IP at maxRequestsPerMinute (default 120, ceiling 200) to stay inside Jumia's stated limits. To go faster, split markets across separate concurrent runs rather than parallelising requests inside one run.

What is the difference between ratingCount and the reviews dataset?

ratingCount is Jumia's count of verified ratings shown on the listing. The Reviews dataset holds individual written reviews. Because ratingCount only counts verified ratings, the Reviews dataset can contain more rows — and more critical reviews — than the headline number suggests.

Can I get product descriptions and specs?

Only from product pages. description and specifications are populated when sourceType is "product"; on listing rows they are null. Category and search walks produce listing rows only, so plan a product-page pass if you need specs.

What does isOfficialStore actually mean?

Whether the seller is Jumia's brand-authorised storefront for that product. It is the cleanest available proxy for first-party pricing versus marketplace reseller pricing — and the single most useful filter for competitive price analysis.

SEO Keywords

jumia scraper, jumia.com, african ecommerce data, jumia api alternative, nigeria ecommerce data, kenya marketplace data, egypt online shopping data, product price monitoring africa, marketplace seller data, official store pricing, discount tracking africa, jumia product data, customer reviews dataset, catalogue ingestion, competitor price analysis, multi country scraping, west africa ecommerce, north africa ecommerce, verified purchase reviews, assortment analysis, online marketplace research, product availability data

For AI Agents & LLM Apps

Returns Jumia product listings across 9 African markets as flat dataset rows, one per product, plus optional customer reviews in a second dataset.

Minimal input — smallest input that produces a valid run:

{ "categoryPaths": ["/phones-tablets/"], "maxItems": 50 }

Multi-market run with compliant identification:

{ "markets": ["ng", "ke", "eg"], "categoryPaths": ["/phones-tablets/"], "botContact": "https://example.com/bot", "maxItems": 300, "maxPagesPerQuery": 3 }

Price monitoring (reviews off to halve cost):

{ "markets": ["ng"], "categoryPaths": ["/laptops-computers-accessories/"], "includeReviews": false, "maxItems": 200, "maxPagesPerQuery": 5 }

Output fields — products (25): country, countryName, currency, title, sku, productId, productUrl, sourceUrl, price, priceFormatted, oldPrice, discountPercent, brand, sellerId, sellerName, isOfficialStore, rating, ratingCount, availability, images, categoryPath, sourceType, description, specifications, scrapedAt

Output fields — reviews (13): country, countryName, sku, productTitle, productUrl, rating, title, text, author, date, isVerifiedPurchase, sourceUrl, scrapedAt

Behaviours an agent should know

• Prefer categoryPaths over searchQueries. Jumia's robots.txt disallows ?q= search URLs; category paths are open. Using searchQueries works but is the disallowed route.
• includeReviews defaults to true, so a default run bills review events too. Set it to false for price-only work — that roughly halves cost.
• Requests are serial from a single IP by design, capped at 200/min (maxRequestsPerMinute, default 120). Do not expect parallelism. To go faster, run markets as separate concurrent runs.
• botContact is baked into the User-Agent. Leave it empty and you are running as a generic bot, which is exactly what Jumia's robots.txt blocks. Set it to a reachable URL or e-mail.
• discountPercent is null when there is no discount (not 0). Do not copy null-handling from the Hepsiburada Actor, which behaves the opposite way.
• description and specifications are only populated on sourceType: "product" rows. Category/search walks give sourceType: "listing" with both null.
• date in reviews is DD-MM-YYYY, not ISO. ratingCount counts verified ratings only.
• One listing page ≈ 40 products. Billing: $0.0009 per product and $0.0009 per review — maxItems alone does not bound cost when reviews are on; bound maxReviewsPerProduct too.
• countries is a legacy alias for markets; prefer markets.

Legal & Compliance

This Actor is an independent third-party tool. It is not affiliated with, endorsed by, or sponsored by Jumia, and it is not associated with any seller, brand or official store whose listings appear in its output.

What it actually accesses: publicly visible Jumia listing, category and product pages, over plain HTTP, with no login, no authentication bypass, no captcha solving and no private endpoints. No headless browser is launched.

Robots policy, handled explicitly. Jumia's robots.txt permits scraping only for a bot that identifies itself with a reachable owner and blocks fake user agents; it caps automated request rates and disallows ?q= search URLs while leaving category paths open. This Actor is built to those terms: botContact is embedded in the User-Agent so the bot identifies itself, requests are serialised from a single IP at or below the 200 requests/minute ceiling, and categoryPaths is the recommended input precisely because that route is open. Set botContact to a real, reachable URL or e-mail before running at any volume.

What that does not mean. Identifying your bot and staying under the rate limit makes automated access tolerated under Jumia's stated policy. It is not a commercial licence. Jumia's Terms of Service govern your use, and scraping the marketplace to build a competing catalogue, price-comparison service or resold dataset may require a separate agreement with Jumia. Use of proxy also deserves a second thought: routing through rotating IPs works around the single-IP design that keeps this Actor within Jumia's limits, so enabling it may put you outside the policy this Actor otherwise respects.

You are solely responsible for compliance with Jumia's Terms of Service, its robots.txt, and applicable law including GDPR and Africa's data-protection regimes. The reviews dataset contains consumer author names attached to public reviews — treat it as personal data, and do not use it to contact reviewers. Keep volumes modest: maxItems (default 100), maxPagesPerQuery (default 1) and maxRequestsPerMinute (default 120) bound a run without effort.