Craigslist Scraper: Listings, Prices & Alerts

Craigslist Scraper: Listings, Prices & Alerts

Craigslist Scraper turns Craigslist postings into clean rows: one row per posting with the title, price, place, coordinates, posted and updated time, category, every image link and, with details (on by default, same price), the full description and every attribute Craigslist shows (cars: make and model, year, odometer, VIN, title status, fuel, transmission; housing: bedrooms, bathrooms, size, laundry, parking, pets; jobs: pay, employment type). Search any of Craigslist's 707 areas and 103 sub areas, whole countries (the United States in six searches), or a ZIP code or coordinates with a radius of up to 1,000 miles, in any of 228 categories, with every Craigslist filter. Paste search links or posting links if you have them. No login, no browser.

Searches larger than Craigslist's 10,000 row limit are split into price slices on their own, and every search ends with a coverage row that says how many postings Craigslist counted and how many were read (a run of very many searches that find little writes the coverage rows past its free row budget as counts in the summary). Name a monitor and schedule the Actor: each run delivers and bills only postings it has not delivered before, tracks price changes, reposts and removed postings, and can send them to Telegram, Slack, Discord, WhatsApp or your webhook.

Who uses it

• Car dealers, flippers and buyers watching cars and trucks by owner or by dealer, with make and model, year, odometer, title status and VIN, across one area or the whole US.
• Landlords, property managers and renters comparing rents, bedrooms, size, pets, laundry and parking in a city or within a radius of a ZIP code.
• Lead and sales teams reading dealer and company postings, which come with their phone numbers and addresses as posted, and ranking postings with a lead score.
• Analysts and developers who need Craigslist as rows in JSON, CSV or Excel through the Apify API, with the same field names in every category and country.

Features

Quick start

1. Pick one or more Craigslist areas (for example sfbay, newyork or houston), or type US in Whole countries.
2. Pick categories (for example cta cars and trucks, apa apartments, jjj jobs; for sale when left empty) and, if you like, keywords.
3. Optional: pick filters (price, sold by owner or dealer, car year and odometer, bedrooms, pets).
4. Set Maximum postings and press Start.
5. Download the rows as JSON, CSV or Excel, or read them through the API.

Example input, the Actor's own prefill (cars and trucks in the San Francisco bay area):

{
    "areas": ["sfbay"],
    "categories": ["cta"],
    "maxItems": 20,
    "proxyConfiguration": { "useApifyProxy": true }
}

Cars by owner in Houston, 2015 or newer, at most 100,000 miles, automatic:

{
    "areas": ["houston"],
    "categories": ["cto"],
    "yearMin": 2015,
    "odometerMax": 100000,
    "transmission": ["automatic"],
    "maxItems": 100
}

Two bedroom apartments in New York at 4,000 or less that take dogs:

{
    "areas": ["newyork"],
    "categories": ["apa"],
    "bedroomsMin": 2,
    "bedroomsMax": 2,
    "dogsOk": true,
    "maxPrice": 4000,
    "maxItems": 100
}

Every Corvette for sale in the United States, search rows only:

{
    "countries": ["US"],
    "categories": ["cta"],
    "keywords": ["corvette"],
    "includeDetails": false,
    "maxItems": 1000
}

Bicycles within 10 miles of a San Francisco ZIP code:

{
    "locations": ["94103"],
    "radiusMiles": 10,
    "categories": ["bia"],
    "maxItems": 100
}

What you get

One row per posting (type: "listing"), one coverage row per search (type: "search"), free status rows for anything that did not end as a posting (type: "status"), with change tracking a removed row per posting a monitor no longer finds (type: "removed"), and one run summary at the end (type: "summary").

Fields and how often Craigslist fills them

Measured on real platform runs of 2026-10-16 (includePersonalData off, the default):

• Search only: 1,000 New York "for sale" rows with details off (run apJfW3J8kLZhF929K; 995 by owner, 1 by dealer, 4 in categories without a seller type).
• Cars with details: 300 Houston "cars & trucks" rows with details (run cBQaUvrLBTa2OUZIe; 93 by owner, 207 by dealer).
• Housing with details: 15 New York apartment rows with details (run iRJvtGJl2KZCJgAYA), filtered to 2 bedrooms, dogs OK and at most 4,000, so bedrooms, price and dogsOk are full by the filter.

Rows with details also carry descriptionHtml (the description as HTML, with descriptionFormat), notices (Craigslist's own notices on the posting), seeMyOther (whether the poster links to their other postings) and createdAt (when Craigslist created the record). A removed row carries type, postingId, a null url, reason, message, lastPrice, firstSeenPrice, priceChangeCount, firstSeenAt, daysTracked, monitorName and scrapedAt; monitors of posting links and runs with a posted window write no removed rows (they do not see every posting of a search).

The lead score adds up plain facts of the row: sold by owner (25), a way to reply (20), a price (15), posted within 24 hours (15), images (10), a description of 80 characters or more (10), not a repost (5).

Sample row (a real row of the prefill run on 2026-10-16; some fields left out)

A private posting, so its link carries no title slug and its coordinates are rounded (see Privacy).

{
    "type": "listing",
    "postingId": "7984331352",
    "postingUuid": "1Sh6LvHnmb9De8zQPmLsna",
    "url": "https://www.craigslist.org/view/d/posting/1Sh6LvHnmb9De8zQPmLsna",
    "title": "2011 Honda Civic Coupe * Gas Saver * Runs Excellent",
    "price": 3950,
    "priceText": "$3,950",
    "currency": "USD",
    "category": "cto",
    "categoryName": "cars & trucks - by owner",
    "sellerType": "owner",
    "area": "sfbay",
    "areaName": "SF bay area",
    "subArea": "eby",
    "subAreaName": "east bay area",
    "region": "CA",
    "country": "US",
    "place": "vallejo / benicia",
    "latitude": 38.11,
    "longitude": -122.24,
    "coordinatesRounded": true,
    "locationType": "pin",
    "postedAt": "2026-10-16T00:56:44.000Z",
    "updatedAt": "2026-10-16T00:56:45.000Z",
    "imageCount": 15,
    "firstImage": "https://images.craigslist.org/00L0L_bpz4YBw1osH_0CI0hT_1200x900.jpg",
    "odometer": 203000,
    "detailsIncluded": true,
    "makeModel": "Honda Civic LX",
    "year": 2011,
    "titleStatus": "clean",
    "fuel": "gas",
    "transmission": "automatic",
    "drive": "fwd",
    "cylinders": "4 cylinders",
    "paintColor": "blue",
    "bodyType": "coupe",
    "condition": "good",
    "repostOf": "7954770517",
    "hasContactInfo": true,
    "streetAddress": null,
    "phoneNumbers": null,
    "emails": null,
    "leadScore": 95,
    "searchLabel": "cta (cars & trucks) in sfbay (SF bay area)",
    "source": "search",
    "missingFields": [],
    "scrapedAt": "2026-10-16T00:57:31.145Z"
}

Coverage row (one per search)

searchKind, searchLabel, searchCategory, searchKeyword, searchPlace, searchRadiusMiles, sort, postingsOnCraigslist (Craigslist's own count), searchCapRows, rowsRead, slices, slicesWalked, countCalls, postingsBeyondReach, callsMade, rowsSeen, distinctPostingsSeen, postingsDelivered, alreadyDelivered, knownToMonitor, filteredOut and filteredOutBy (per filter), nearbyAreaRows, deletedOrGone, detailsFailed, newPostingsMayLieBeyond (a monitor), complete, stoppedBy, moreRowsOffered and a note.

Removed rows (trackChanges)

postingId, reason (notFoundByCompleteSearches), message, lastPrice, firstSeenPrice, priceChangeCount, firstSeenAt, daysTracked, monitorName, scrapedAt.

Free status rows

A setting or an item the Actor refused (invalid), an area or category name that matches several entries (ambiguous, with the candidates), a posting link Craigslist does not know or a search it does not serve (not_found), a posting deleted by its author (deleted), a posting or a search that could not be read (failed), a search whose filters Craigslist did not apply as sent (search_not_applied), a posting link a monitor already delivered (skipped), a notification that failed (notification_failed), a run the platform restarted (resumed), and the reason a run stopped (limit_reached, source_unavailable, monitor_unavailable). A run writes 20 such rows plus one per billed event, at most 200 of one kind; past that it counts the rest in the summary. Coverage rows share that budget.

How searches are read

• Past 10,000 rows. Craigslist serves at most the newest 10,000 rows of one search. When a search holds more and your caps want more, it is split into whole dollar price slices (each planned with a call of 360 rows, halved while a slice holds more than 9,000), and each slice is read from the top. Postings without a price fall in no slice: they are reached only among the newest 10,000, and the coverage row counts every posting left beyond reach. In a test of 2026-10-16 the San Francisco bay area cars search counted 15,588 postings and was read in 2 price slices.
• Every filter checked. Craigslist ignores a parameter it does not know and answers an option value outside its list with an empty page, without an error. The Actor sends only Craigslist's own filter names and values, refuses a pasted link with any other, and compares Craigslist's echo of the search with what it sent on every answer; a search that differs ends with a free search_not_applied row and nothing billed from that answer. Every row is also checked against the filters its own fields show (area, sub area, country, price, seller type, images, the car and housing attributes), again on the posting when details are read, and a posting edited between the search and the read is checked against the keyword again.
• Nearby areas. When a search has few local postings, Craigslist adds postings of nearby areas after them and does not count them. They are left out free unless you turn on includeNearbyAreas, and delivered ones carry isNearby: true.
• The whole US. Craigslist has no country search. countries: ["US"] runs six searches of 1,000 miles (Craigslist's largest radius) around Louisville, Provo, New Orleans, Detroit, Anchorage and Honolulu, which hold every US area within 800 miles of a centre; a posting found by two circles is delivered once, and postings of Canadian or Mexican areas inside a circle are left out free.
• One row per posting. A posting found by several searches in one run is delivered and billed once.
• Fresh answers. Craigslist keeps each search answer for 15 minutes for every client; the Actor asks for a fresh one on every call. In a test of 2026-10-15 the newest "for sale" posting was 17 to 49 seconds old in three cities.
• Access. Requests leave through one Apify datacenter exit per run (a new one only when the proxy itself fails; when the proxy's address cannot be checked three times, or every session lands in a network Craigslist refuses, the run reads from its own container's address and the summary says so), two at a time, 2 a second by default, and only through doors Craigslist's robots.txt allows. A refusal is asked once more on the same exit after a pause; a second refusal ends the run's reading with a free source_unavailable row. No captcha is ever solved.

How much does it cost to scrape Craigslist?

You pay per posting row delivered, with the full details included, plus a small start fee per run. Platform usage and the datacenter proxy are included.

Platinum and Diamond plans pay the Gold price. The Actor start event is a flat $0.007, charged once per run, with no plan discount. The Store pricing card shows these same prices per 1,000 events: "$0.60 / 1,000" on the Listing row means one posting costs $0.0006, and "$0.64 / 1,000" on the Results page row means each 360 rows a monitor checks cost $0.00064.

• 1,000 postings with details cost $0.66 + $0.007 = $0.667 on Free, $0.60 + $0.007 = $0.607 on Bronze, $0.55 + $0.007 = $0.557 on Silver and $0.47 + $0.007 = $0.477 on Gold.
• The prefill run (20 postings with details) costs 20 x $0.00066 + $0.007 = $0.0202 on Free and 20 x $0.00047 + $0.007 = $0.0164 on Gold.
• An hourly monitor over one search on Bronze reads the newest 360 rows each run: 24 runs a day cost 24 x ($0.007 + $0.00064) = $0.18336 a day, plus $0.0006 per new posting. A run that finds 12 new postings costs $0.007 + $0.00064 + 12 x $0.0006 = $0.01484.
• What is never billed: postings left out by your filters, nearby area postings left out, a posting found twice in one run, postings a monitor already delivered (unless change tracking finds a change, or you ask for every known posting), deleted and unknown postings, a search Craigslist does not serve or whose filters it did not apply, refused inputs, search calls and posting reads themselves (outside a monitor; in a monitor run each search call is a Results page), coverage, status and summary rows, the run report and the notifications.
• Limits: set Maximum postings (100 when left empty) and, in the run options, a maximum charge for the run: the Actor stops before it would pass either, and a row whose events do not all fit is withheld whole. Maximum postings counts postings, changes and removed rows together. A run also spends a small allowance on work that delivers nothing billable (searches whose rows your filters leave out, failed reads), USD 0.0008 plus a share of every paid event it delivers; past it the run stops with a free row that says so.

Input

A run holds up to 300 searches (places x categories x keywords x days, plus one per search link) and up to 5,000 posting links. A filter that does not fit a searched category (bedrooms on cars, for example) refuses the run with a free row.

The monitor and notifications

Name a monitor (monitorName) and schedule the Actor (every hour, for example). The monitor remembers every posting it delivered in a named store this Actor creates for itself, up to 60,000 postings, so each run delivers and bills only postings it never delivered before; known postings are skipped for free. A monitor reads the newest 360 rows of each search by default (set maxRowsPerSearch to read 1,080, 3,240 or 10,000), does not split searches, and pays a Results page for each 360 rows it reads (at least one per search); when the coverage row says newPostingsMayLieBeyond, read more rows or run more often. One monitor name belongs to one set of settings: a run with other settings delivers nothing and gets a free row saying so.

• Change tracking (trackChanges): a known posting whose price changed, or whose price was taken away or put back, comes back with change (changeType: "priceChanged", previous price, difference, first seen price, price history, drop in percent since first seen), billed as a Listing update. With trackFields a change of the title, the images, the location (place, neighbourhood, map point), the odometer, bedrooms, square feet, pay, monthly payment or the sale, event or open house dates counts too (changeType: "updated", changedFields): the facts every search row shows, compared on it, so tracking them reads nothing more. A change of the description or the attributes alone is not tracked: it would need every known posting read again on every run. A posting the monitor reported as removed that shows up again is reappeared.
• Removed postings: when every search of the run was read to its end (for a monitor, a search that holds no more postings than its window), a posting the monitor tracks that the run did not meet gets a removed row. When more than half of the tracked postings (and more than 5) are missing at once, none are reported and the run says so in a free row.
• Snapshots and reposts: emitUnchanged delivers known postings again (billed as Listing); markReposts marks a new posting that reposts one the monitor delivered (Craigslist's own "repost of" mark, or the same area, category, title and price under a new id) with possibleRepostOf, and skipReposts leaves it out free. Both need details.
• Notifications (free, with or without a monitor) after the rows were delivered and billed: Telegram (your own bot, one message per posting), Slack and Discord incoming webhooks and WhatsApp Cloud API (one message per run), each naming up to notificationLimit postings (10 by default, up to 30); a webhook (webhookUrl) gets JSON batches of 100 postings in a short form (id, link, title, price, currency, price text, place, area, category, seller type, miles, bedrooms, posted time, image, search label, change), up to 1,000 a run, signed in the header x-signature-sha256 with webhookSecret, with your own webhookHeaders. With a monitor, unchanged postings are left out of the messages (notifyOnlyChanges, on by default). A failed call never fails the run: it is a free notification_failed row, without your address or token.

A feed of new postings (in place of Craigslist RSS)

Craigslist's RSS feeds refuse automated readers, so this Actor never uses them. To get a stream of new postings:

1. Pick your search (areas, categories, keywords, filters) and give it a monitorName, such as sfbay-tacoma.
2. Add a webhookUrl (Zapier, Make, n8n or your own endpoint), or a Telegram, Slack, Discord or WhatsApp channel.
3. Save the input as a task and schedule it in Apify (every 15 minutes or every hour).

Each run then sends only postings the monitor has not seen. If a tool of yours reads RSS, Apify's API serves any dataset in RSS format (format=rss on the dataset items endpoint); the feed then holds the run's coverage, status and summary rows next to the postings.

Past the 10,000 row cap

Leave splitLargeSearches on (the default) and set maxItems to what you want: a search that Craigslist counts above 10,000 is read slice by slice by price, each slice from the top, and the coverage row gives Craigslist's count, the slices read and anything left beyond reach. A monitor does not split; give it narrower searches (a sub area, a category, a price range) instead.

Privacy and personal data

Many Craigslist postings come from private persons. With includePersonalData off (the default), every posting that is not in a "by dealer" category and names no company is treated as private: phone numbers, e-mail addresses, street addresses written with a house number and a street word (4512 Main St, Hauptstraße 12, Calle Mayor 12) or with a house number and a state code (123 Broadway, New York, NY 10001), exact coordinates written into the text (37.123456, -122.654321, in either order, also inside a map link), and names written right after a contact word (in lower case, the common first names) (

ask for John

Legal

This Actor is not affiliated with or endorsed by Craigslist. Craigslist is a trademark of its owner. It is an unofficial tool that reads public postings without logging in, through the doors Craigslist's robots.txt allows. Craigslist's terms of use forbid collecting its content with scrapers, crawlers or scripts. You are responsible for a lawful use of the data, for complying with privacy laws for any personal data, and for keeping a scheduled monitor to what you need.

What this Actor does NOT do

• No reply e-mail or phone number from behind the reply button. Craigslist shows the reply contact only after a captcha; the Actor never opens it. hasContactInfo says whether a posting offers one; the posting's link (url) is where a buyer replies.
• No tracking of description or attribute changes alone. Change tracking compares the facts every search row shows; the description and the attributes would need every known posting read again on every run.
• No posting, flagging, messaging, saved searches or anything else that needs a Craigslist account. No login.
• No private poster's phone numbers, e-mail addresses, street address field or exact coordinates unless you turn on includePersonalData. Addresses and names written in the texts are masked in the forms Privacy names. Dealers' and companies' contacts are always delivered.
• No captcha solving and no new exit to get past a refusal. A second refusal in a row ends the run's reading with a free source_unavailable row.
• No RSS feeds and no sitemaps. Searches go through Craigslist's search, postings through its posting pages.
• No more than 10,000 rows from one Craigslist search answer. Past that the search is split by price (not in a monitor); postings without a price are reached only among the newest 10,000, and the coverage row counts what is left.
• No postal codes outside the US. Give coordinates for a radius search elsewhere, or the areas themselves.
• No view counts. Craigslist shows none, and reading a posting counts nothing.
• No image downloads. Images are links in the size you pick.
• No translation. language sets Craigslist's labels; descriptions come as the poster wrote them.
• No export into Notion, Airtable or other apps from inside the run. Those exports need your own account at each app; connect the run's dataset with Apify's integrations or MCP connectors, authorised in your own Apify settings.
• No preloaded snapshot. Every run reads Craigslist live; the data is as fresh as the run.

FAQ

Does Craigslist have an API?

Craigslist offers no API to read or search its postings: its official interfaces serve the list of areas and categories and the bulk posting of ads. This Actor reads the same public postings any visitor sees, with no login, and gives them to you as rows you can read through the Apify API (examples below).

Is it legal to scrape Craigslist?

This Actor reads only public postings without logging in, never uses the reply button, flagging or posting, and masks the phone numbers, e-mail addresses and street addresses it finds in private postings unless you turn personal data on. Craigslist's terms of use forbid collecting its content with scrapers, so you are responsible for a lawful use of the data, privacy laws included (see Legal).

How do I search all of Craigslist at once?

Put US in Whole countries: the Actor runs six searches of 1,000 miles that together cover every US area, delivers each posting once and leaves out postings of other countries. Other countries are searched area by area, and a ZIP code or coordinates with a radius of up to 1,000 miles searches every area around a point.

How do I get alerts for new Craigslist listings?

Name a monitor and schedule the Actor, every hour for example. Each run delivers and bills only postings the monitor has not delivered before, and can send them to Telegram, Slack, Discord, WhatsApp or your webhook; no Craigslist account is needed.

What can I use instead of Craigslist RSS feeds?

Craigslist refuses its RSS feeds to automated readers (HTTP 403 in our tests), so this Actor never uses them. A scheduled monitor with a webhook gives a stream of new postings, and Apify's API can serve a run's dataset in RSS format for a feed reader (see A feed of new postings).

Why does this Actor run with limited permissions?

It needs nothing beyond its own storages: it reads Craigslist's public postings and writes to its own dataset and key-value store, plus the named store of a monitor you set with monitorName, which the Actor creates for itself. Limited permissions keep it from touching anything else in your account.

API

Each example starts a run and waits for its rows (the run-sync endpoint suits short runs; use maxItems to keep them short).

curl -X POST "https://api.apify.com/v2/acts/automation_craft~craigslist-scraper/run-sync-get-dataset-items" \
  -H "Authorization: Bearer <YOUR_API_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"areas": ["sfbay"], "categories": ["cta"], "maxItems": 20}'

const res = await fetch('https://api.apify.com/v2/acts/automation_craft~craigslist-scraper/run-sync-get-dataset-items', {
    method: 'POST',
    headers: { Authorization: 'Bearer <YOUR_API_TOKEN>', 'Content-Type': 'application/json' },
    body: JSON.stringify({ areas: ['sfbay'], categories: ['cta'], maxItems: 20 }),
});
const rows = await res.json();
console.log(rows.filter((r) => r.type === 'listing'));

import requests

res = requests.post(
    "https://api.apify.com/v2/acts/automation_craft~craigslist-scraper/run-sync-get-dataset-items",
    headers={"Authorization": "Bearer <YOUR_API_TOKEN>"},
    json={"areas": ["sfbay"], "categories": ["cta"], "maxItems": 20},
    timeout=330,
)
postings = [r for r in res.json() if r["type"] == "listing"]

Changelog

See the Changelog tab. 0.1: first release.

More data tools by Automation Craft

• OfferUp Scraper: Local Listings, Prices, Sellers
• Facebook Marketplace Scraper with Sold Tracker
• Kijiji Scraper: Listings, Cars, Rentals & Prices
• Mercari US Scraper: Listings, Sold Prices
• Kleinanzeigen Scraper & API: Ads, Prices, Places
• Leboncoin Scraper: Ads, Prices, Real Estate & Cars
• AutoScout24 Scraper: EU Car Listings & Prices
• Mobile.de Scraper: Cars, Prices, Dealers & Ratings