Kijiji Real-Time Data

Live resale market intelligence for Kijiji — thrift price comps, seller inventory, vintage deal finder. Keyword search, full listing details, seller tracking, sold comps, and URL collection. Checkbox features — enable only what you need. Structured JSON streamed to your dataset in real time.

💎 Paid plans get the full experience. Free Apify plans are capped at 2 results per run so you can try the actor before upgrading. Upgrade to any paid Apify plan for unlimited vintage and thrift data exports.

What it does

Kijiji is Canada's largest classifieds marketplace — and one of the most underrated sourcing channels for thrift and vintage resellers. Every day, private sellers list vintage Carhartt, single-stitch band tees, 90s Levi's, gorpcore shells, and underpriced designer pieces for far below their resale value.

This actor gives you structured, real-time access to that market:

• Find underpriced inventory fast — live keyword search with price, seller-type, and photo filters
• Pull full product details — descriptions, seller-filled attributes (brand, size, condition), complete photo sets, view counts
• Track sellers — a seller's entire active inventory, their profile stats, ratings, and verification status
• Comp your prices — ended and removed listings per seller for margin and sell-through analysis
• Automate everything — Slack/Discord deal alerts via webhooks, or pipe rows into Airtable, Sheets, or your pricing bot

Who it's for

• Vintage & thrift resellers cross-listing on Poshmark, eBay, Depop, Mercari, and Grailed who source from Canadian classifieds
• Resale arbitrageurs hunting undervalued vintage single-stitch, designer denim, gorpcore, and Y2K pieces
• Consignment & vintage shop owners monitoring local inventory and pricing
• Pricing teams & market researchers building multi-platform comps datasets
• AI agents & MCP workflows that need clean, structured resale data

Feature matrix

Feature	What you get	Output featureType
🔍 Listing Search	Search-task list: each row pairs a keyword with an optional location (city/region/province) and category, plus price / seller-type / photo filters and sort order. On by default.	listing_search
📦 Listing Details	Full record for specific listings by URL or ID: description, all seller-filled attributes, full gallery, views, status.	listing_details
👚 Seller Inventory	A seller's complete active listing inventory — track competitor stock and new drops.	closet_listings
👤 Seller Profile	Display name, seller type (private/dealer), inventory size, verification, member since, review summary.	seller_profile
🧾 Sold Item Comps	Ended and removed listings per seller — the comp data behind pricing decisions.	sold_history
🔗 Scrape By URL	Paste any Kijiji URL (search, category, seller, single listing) — page type is auto-detected.	scrape_by_url

Enrichment, not duplication: when "Enrich with full listing details" is enabled on Listing Search, each search card is upgraded in place — you still get exactly one row per listing (with detailsFetched: true), just with the full description, attributes, and gallery merged in. No duplicate rows, nothing filtered out.

Tags, not filters — predictable run volume: the actor enables details instead of filtering by them. Every listing it discovers goes to the dataset, even if some fields are empty. Freshness (alert mode), category, and location are the only things that scope which listings are searched — never how many rows come back. If you ask for 100 results per keyword, you get 100 rows. That makes per-run volume (and therefore cost) fully predictable: a 1,000-listing run is exactly 1,000 dataset rows, every time. Filtering happens on your side, downstream, on the tag columns (is_newly_listed, posted_age_days, category_path, …).

Quick start

1. Click Start — Listing Search is on by default with three demo search tasks pre-filled (keyword + location / category pairs).
2. Within seconds you'll have structured results in your dataset — one row per listing discovered.
3. Edit the Search tasks list: one row per search, pairing a keyword with an optional place (e.g. City of Toronto) and category (e.g. men's shoes) — like a maps-style keyword + location search.
4. Open the dataset table to browse, or grab the JSON/CSV/Excel export.

Practical search examples

Each row of the Search tasks list is one search — a keyword, optionally pinned to a place and a category:

Goal	Keyword	Location	Category
Vintage workwear flips in Toronto	carhartt detroit jacket	Toronto (GTA)	—
Denim sourcing province-wide	vintage 90s levis 501	Ontario	—
Sneakers, men's only	nike jordan	—	men's shoes
Band tees in Vancouver	single stitch band tee	Vancouver	—
Anything free, anywhere	free	—	—

Feature-wide filters (price, seller type, photos, sort) in the section below apply to every task.

Location targeting

Like a maps-style search, you combine a keyword with a place. The Location column of a search task accepts city, region, or province names — or a numeric location id:

• Provinces: Ontario, Quebec, Alberta, British Columbia
• Regions: Toronto (GTA), Kitchener Area, Windsor Region, Muskoka
• Cities: City of Toronto, London, Barrie, Vancouver

The first match wins per task; matching walks the marketplace's official location tree (Canada → provinces → regions → cities), so names resolve to the same ids the site itself uses. Leave the Location column empty to search all of Canada — and if a name can't be matched, the task falls back to all of Canada with a warning in the log. Every row still carries the seller's own city in seller_location.

Input reference

🔍 Listing Search

Field	Type	Default	Notes
Listing Search	checkbox	true	Master switch for keyword search
Search tasks	task list	3 demo tasks	One row per search: Keyword (required) + optional Location, Category, and per-task Max results. This — not filtering — is what sets your run volume
Default max results per task	integer	10	1–200; used when a task omits Max results
Category scope / Location for all tasks	string list	empty	Advanced: apply one scope to every task; multiple selections are searched any-of. Prefer the per-task columns
Sort order	select	relevance	relevance, newest (best for deal alerts), price_low_to_high, price_high_to_low

🔍 Search keywords (alternative)

Simple mode without the task table: a plain keyword list, searched everywhere with no per-task location/category. Only used when the Search tasks list is empty.

Field	Type	Default	Notes
Search keywords	string list	empty	e.g. carhartt detroit jacket, vintage 90s levis 501

Feature-wide search filters (apply to whichever search mode above provided the keywords):

Field	Type	Default	Notes
Min / Max price	integer	empty	CAD
Price type	select	Any	Fixed, Contact seller, Give away, Swap/trade
Alert mode (newest first)	checkbox	off	Sorts newest-first and tags every row with is_newly_listed + posted_age_days. Nothing is filtered out — run volume stays exactly what you set
Freshness window (days)	integer	7	Listings posted within this window are flagged is_newly_listed — a tag for downstream filtering, not a row filter
Enrich with full listing details	checkbox	off	Upgrades each search row in place with full details (adds a little extra time per listing)

Tags, not filters: owner-vs-dealer and photo completeness are not run options — every row carries is_private_seller (true = private seller, false = dealer) and has_image so you filter downstream, in your spreadsheet or code. This keeps run volume exactly what you asked for: a 1,000-listing run costs 1,000 result events and returns 1,000 rows, every time, regardless of how many happen to be dealer rows or photo-less.

📦 Listing Details

Field	Type	Notes
Listing IDs	string list	The long number at the end of a listing URL, e.g. 1737865031
Listing URLs	string list	Full https://www.kijiji.ca/v-... URLs

👚 Seller Inventory · 👤 Seller Profile · 🧾 Sold Item Comps

Field	Type	Default	Notes
Seller IDs	string list	—	The seller's numeric ID — from a profile URL like kijiji.ca/o-profile/1015260943/.... Shared by all three features
Max listings per seller	integer	30	Seller Inventory, 1–500
Max items per seller	integer	30	Sold Comps, 1–500

🔗 Scrape By URL

Field	Type	Notes
Kijiji URLs	string list	Search results, category pages, seller profiles, or single listings. Page type is auto-detected; non-Kijiji URLs are rejected

🔔 Delivery

Field	Type	Default	Notes
Webhook URL	string	empty	Every record is always saved to the dataset; this is an additional real-time POST per record (Slack, Discord, Zapier, Make, n8n, custom bots)
Webhook format	select	json	json (full record) or slack (formatted message)

🌐 Connection

Field	Notes
Proxy settings	Residential proxy with CA (Canada) exits is strongly recommended — it is the most reliable setup for this Canadian marketplace. Defaults are preconfigured.

Output field reference

Every row carries featureType (which feature produced it), scrapedAt, and url. Filter your dataset view by featureType.

Core listing fields (search, details, inventory, comps, URL scrape)

Field	Type	Description
item_id	string	Native listing ID (stable, use for dedupe)
item_url	string	Canonical listing URL
title	string	Listing title
description	string	Seller-written description (search cards get a short excerpt; full text on enriched/details rows)
main_image_url	string	Cover photo
additional_image_urls	string[]	Full photo gallery
status	string	available, sold, reserved, …
current_price	number	Asking price (major units, e.g. 150.00)
original_price	number	Seller's stated original price, when set
currency	string	CAD
discount_percentage	number	Markdown vs. original price
price_drop_amount	number	Absolute price drop
price_type	string	fixed, contact_seller, give_away, swap_trade, make_offer
brand	string	From seller-filled attributes (Carhartt, Levi's, Patagonia, …)
size	string	Seller-filled size
condition	string	Seller-filled condition
category / department / subcategory	string	Catalog classification where available
category_id	string	Numeric catalog category id — stable key for category-scoped sourcing
category_path	string	Full catalog path, e.g. Clothing > Women's - Tops & Outerwear
color / material / style_tags	string / string[]	Seller-filled style attributes
created_at	string	When the ad was posted
updated_at	string	Last bump / sort timestamp
is_sold	boolean	Sold/ended flag
is_price_dropped	boolean	Seller used a price drop
is_newly_listed	boolean	Posted within the freshness window (default 7 days) — a tag, never a row filter
posted_age_days	number	Whole days since the ad was posted, so you can slice freshness downstream
seller_username / seller_id	string	Seller's numeric ID (the stable identity on this marketplace)
seller_verified	boolean	Platform-verified seller
seller_location	string	Listing location (city/province)
position	number	Rank on the page it was found on
detailsFetched	boolean	Search rows only: true when enriched in place

Listing Details extras

views_count (view counter), item_specifics (every seller-filled attribute group), image_count, end_date, has_video, is_featured, is_top_ad, is_highlight, seller_type.

Seller Inventory (closet_listings)

One row per active item: core pricing/product fields + seller_username, seller_id, location, created_at.

Seller Profile (seller_profile)

Field	Description
display_name	Seller's display name
seller_type	FSBO (private) or commercial
total_listings	Active inventory size
is_verified	Verification status
member_since	Account age
rating / total_reviews	Review summary
profile_url	Canonical profile URL

Sold Item Comps (sold_history)

One row per ended/removed listing with pricing, product fields, status, and sold_at where determinable. Use for margin analysis and sell-through.

Scrape By URL (scrape_by_url)

Adds pageType (listing, seller, search, category) to the row so you can segment bulk URL runs.

Webhooks & deal alerts

Set Webhook URL and every record is POSTed the moment it's collected — perfect for Slack/Discord deal bots. Turn on Alert mode to sort results newest-first so the freshest inventory arrives first; every row still lands in the dataset (tagged, not filtered), so your alert volume stays predictable.

• json format — the full record object. Ideal for Zapier, Make, n8n, or your own automation.
• slack format — a ready-to-post Slack message with title, price, brand, seller, and a link.

{
  "featureType": "listing_search",
  "title": "Vintage Carhartt Detroit Jacket",
  "current_price": 150,
  "currency": "CAD",
  "brand": "Carhartt",
  "item_url": "https://www.kijiji.ca/v-...",
  "is_newly_listed": true
}

Webhook delivery is best-effort and never blocks or fails the run — the dataset write always happens first.

Using with AI agents (MCP)

This actor is Apify MCP-compatible. Point your agent at it and ask things like:

• "What's the average asking price for a 90s Carhartt J97 jacket on Kijiji right now?"
• "Find vintage Levi's 501 listings under $60 posted this week and alert me on Slack."
• "Track seller 1015260943's inventory and tell me when they list new denim."
• "How fast do vintage band tees sell for sellers in Toronto?"

Free tier & paid plans

Plan	What you get
Free	2 results per run — enough to validate data quality and see the output structure
Any paid plan	Unlimited results, every feature, full enrichment

Free-tier runs exit gracefully with a clear message (never an error), and the paywall object in the run OUTPUT discloses exactly what was applied. Owners can configure the behavior with Console environment variables (FREE_TIER_MODE = limit/block, FREE_TIER_MAX_ITEMS).

FAQ

How fresh is the data? Every run hits the live marketplace — results reflect what's listed at that moment. Use "Just listed" sort for deal alerts on fresh inventory.

Do I need proxies? Residential proxies (CA) are configured by default and recommended for reliability. This is a Canadian marketplace, so Canadian exit addresses work best.

Can I get sold prices? Sellers curate their public inventory and ads leave the feed once ended or sold. Sold Comps collects each seller's ended/removed listings alongside active ones — the closest available comp signal on this platform.

One row or two per listing? Always one. Enrichment merges into the same search row (detailsFetched: true) — never a duplicate.

Rate limits? The actor paces requests with small random pauses, reuses warmed sessions, and retries transient failures automatically. Keep per-run volume sensible (hundreds, not tens of thousands, of enrichments) and split very large jobs into runs.

Is the seller ID the same as a username? This marketplace uses numeric seller IDs (visible in profile URLs) rather than public usernames. That ID is the stable key for Seller Inventory, Seller Profile, and Sold Comps.

Pricing models & monetization

This actor uses Pay-Per-Event — you're charged per result event (result), so a run that finds nothing costs nothing beyond platform usage. Set your max run cost in Apify Console to control spend; the actor respects spending limits and stops gracefully when the budget is reached.