TechCrunch Scraper: full-text tech news for analysts, RAG and MCP agents

TechCrunch Scraper returns structured, full-text TechCrunch articles as clean markdown, plain text or HTML. You can search by keyword, follow a category, author or tag, limit by date range, or enrich a list of article URLs you already have. It covers tech news, startup funding news, AI, fintech and venture, and the output drops straight into a spreadsheet, a vector database (RAG) or an AI agent through Apify's MCP server.

What does TechCrunch Scraper do?

• Keyword search. Each keyword runs as its own TechCrunch search, for example "anthropic" or "series b".
• Category, author and tag feeds. Follow artificial-intelligence, fintech or venture, a reporter, or a tag like openai.
• Date ranges. Use absolute dates (2026-10-01) or relative ones (7 days, 2 weeks, 1 month, yesterday).
• URL list enrichment. Paste TechCrunch article links and get back the full article with metadata.
• Only new articles on scheduled runs. Turn on dedupe and each run returns only articles that earlier runs haven't returned.

Each result includes the title, URL, authors, publish and update dates, category and tag names, excerpt, featured image, word count, reading time, paywall and sponsored flags, and the full body text.

Why use this TechCrunch scraper?

• Full text, ready for LLMs. You get the body in markdown, text or cleaned HTML, with ads, newsletter prompts and related-link blocks removed.
• Exact-phrase keyword matching. Raw TechCrunch search returns loose matches. Results are re-checked so "series b" returns only articles that actually say "Series B".
• No self-promo noise. TechCrunch's own event ads (posts under the "TechCrunch Events" byline, about 10% of the Startups feed in October 2026) are skipped by default. Ticket call-to-action lines inside news stories are stripped too.
• Every filter combines with the others. You can mix keywords, categories, tags, authors, dates and URLs in a single run.
• Precise date ranges for backfills or tracking a company's coverage.
• Multi-author support. Every co-author is listed, and author stays flat for CSV users.
• Tags and categories come as human-readable names, not numeric ids.
• New-only scheduled runs with cross-run dedupe.
• No browser. The actor reads TechCrunch's public WordPress API directly, so it's fast and cheap per article.

Input modes

You can use any combination of the fields. Leave everything empty to get the latest TechCrunch articles.

1. Keyword search

{
    "keywords": ["anthropic", "series b"],
    "dateFrom": "30 days",
    "maxItems": 50,
    "bodyFormat": "markdown"
}

2. Category / author / tag feed

{
    "categories": ["artificial-intelligence", "venture"],
    "authors": ["kyle-wiggers"],
    "dateFrom": "2026-09-01",
    "dateTo": "2026-10-25",
    "maxItems": 200
}

3. Article URL list

{
    "startUrls": [
        "https://techcrunch.com/2025/01/27/deepseek-displaces-chatgpt-as-the-app-stores-top-app/",
        "https://techcrunch.com/2024/12/20/openai-announces-new-o3-model/"
    ],
    "bodyFormat": "text"
}

Scheduled daily feed with only new articles

Create a schedule (for example daily at 07:00) with this input. Each run returns only articles that earlier runs with the same dedupeStoreName haven't returned.

{
    "categories": ["fintech"],
    "dateFrom": "3 days",
    "maxItems": 200,
    "dedupe": true,
    "dedupeStoreName": "techcrunch-fintech-daily"
}

Use a different dedupeStoreName for each feed. If you delete the store, its memory is reset.

How the inputs combine

Input	Behavior
keywords	Each keyword runs as a separate search. Results are merged.
strictKeywordMatch	On by default. TechCrunch's own search matches the words separately, so series b would also return an article about "a series of apps". With this on, only articles that contain the exact phrase in the title, excerpt or body are kept. Dropped results don't count toward maxItems.
categories, tags, authors	Filters: OR within a field and AND across fields. For example, categories: [ai, fintech] + authors: [x] means "(AI or Fintech) and by x". They apply to every keyword search.
Filters without keywords	One feed query runs using the filters.
Nothing at all	The latest articles.
startUrls	Processed in addition to any search, before the search results.
includeEventPromos	Off by default, which skips TechCrunch's event marketing posts. Start URLs are never skipped.
dateFrom / dateTo	Inclusive. A date-only dateTo covers the whole day (UTC).
maxItems	Total cap across all sources, after deduplication.

Results are deduplicated within a run by post id and canonical URL, so tracking parameters, www. and /amp/ variants count as the same article. Each item's matchedBy lists every source that found it, for example ["keyword:anthropic", "url"]. Feed results are labeled with the specific filter values they matched, for example category:fintech or author:kyle-wiggers. With several sources, results are sorted newest first.

If a category, tag or author can't be found, the run logs a warning with suggestions and continues with the values that did resolve. If none of the values for a field resolve, searches using that filter return no results. The actor never silently drops the filter and returns everything.

Categories, tags and authors accept a slug (artificial-intelligence), a display name (AI), or a TechCrunch URL (https://techcrunch.com/author/kyle-wiggers/).

Sample output

{
    "id": 3179660,
    "url": "https://techcrunch.com/2026/10/26/matiks-raises-10m-to-expand-its-math-and-puzzle-gaming-platform/",
    "slug": "matiks-raises-10m-to-expand-its-math-and-puzzle-gaming-platform",
    "title": "Matiks raises $10M to expand its math and puzzle gaming platform",
    "excerpt": "Matiks has introduced a new platform for users to create games using AI.",
    "author": "Ivan Mehta",
    "authorUrl": "https://techcrunch.com/author/ivan-mehta/",
    "authors": [
        {
            "name": "Ivan Mehta",
            "slug": "ivan-mehta",
            "url": "https://techcrunch.com/author/ivan-mehta/"
        }
    ],
    "publishedAt": "2026-10-26T14:08:04Z",
    "updatedAt": "2026-10-26T16:02:09Z",
    "categories": [
        "Apps",
        "Fundraising"
    ],
    "primaryCategory": "Apps",
    "tags": [
        "games",
        "General Catalyst"
    ],
    "body": "Over the past decade, head-to-head brain games like QuizUp rocketed in popularity, reaching at one point millions of users. After that, battle royale games like Fortnite and PlayerUnknown’s Battlegrounds had their moment in the spotlight. Lately, growth of [mobile games has slowed …",
    "bodyFormat": "markdown",
    "imageUrl": "https://techcrunch.com/wp-content/uploads/2026/10/Matiks-co-founders.jpg",
    "wordCount": 673,
    "readingTimeMinutes": 3,
    "isPaywalled": false,
    "isSponsored": false,
    "language": "en",
    "matchedBy": [
        "keyword:series a"
    ],
    "source": "api",
    "scrapedAt": "2026-10-27T03:40:21.926Z"
}

This is a real item from a run on Apify (October 2026), with body shortened for display.

Output fields

Field	Type	Description
id	number or null	WordPress post id (null for items parsed from the HTML page)
url	string	Canonical article URL
slug	string	URL slug
title	string	Headline
excerpt	string or null	Dek or summary
author	string or null	Primary author's name
authorUrl	string or null	Primary author's profile page
authors	object[]	All authors: name, slug, url
publishedAt	ISO 8601 or null	Publish time (UTC)
updatedAt	ISO 8601 or null	Last update (UTC)
categories	string[]	Category names
primaryCategory	string or null	Main category
tags	string[]	Tag names
body	string or null	Full text in bodyFormat; null when includeBody is false
bodyFormat	string or null	markdown, text or html
imageUrl	string or null	Featured image
wordCount	number	Words in the body
readingTimeMinutes	number	Estimated reading time
isPaywalled	boolean	True when the public body is truncated (TechCrunch+ or a gate)
isSponsored	boolean	Sponsored or partner content
language	string	Content language (en)
matchedBy	string[]	Sources that found the article: keyword:…, category:…, tag:…, author:…, url, latest
source	string	api (WordPress API) or html (page-parser fallback)
scrapedAt	ISO 8601	When the item was scraped

You can export the dataset as JSON, CSV, Excel, XML or RSS from the Output tab or through the API. The STATS record in the run's key-value store has counts for that run: found, pushed, duplicates skipped, previously seen skipped, and failed URLs.

Pricing

The actor uses pay per event:

Event	Price	When it's charged
Actor start	$0.00005 per run	Once when a run starts.
Article	$0.002 per article ($2.00 per 1,000)	Once for each article saved to the dataset.

A run that finds nothing costs only the $0.00005 start fee. You aren't charged separately per dataset result, only per article. Set a maximum charge per run in the run options. The actor stops cleanly when that limit is reached.

Use with MCP and AI agents

The actor is MCP-ready: input names are standard, no field is required, and every field has a description and an example. To add it to Claude Desktop, Cursor or another MCP client through Apify's MCP server, select the actor scrapelabmax/techcrunch-scraper. An agent can then answer "what did TechCrunch say about Anthropic this week?" with a call like:

{ "keywords": ["anthropic"], "dateFrom": "7 days", "maxItems": 20 }

For RAG ingestion, use bodyFormat: "markdown", chunk the body field, and store url, title, publishedAt, categories and tags as metadata.

FAQ and limits

Does it need a proxy? No. The TechCrunch API works without one. If you hit rate limits, enable Apify datacenter proxy in the Advanced section.

What about paywalled articles? The actor returns only what is publicly available. If the public body is truncated, isPaywalled is true, so the item is never passed off as a complete article.

Can I scrape thousands of historical articles? Yes. Use dateFrom and dateTo with a high maxItems (up to 100,000). Results come newest first.

Why was one of my URLs skipped? Only TechCrunch article URLs (techcrunch.com/YYYY/MM/DD/slug/) are processed. Category pages, author pages, the homepage and other sites are skipped with a warning in the log.

How does dedupe work? With dedupe: true, the actor stores a SHA-1 hash of each returned article's canonical URL in a named key-value store, and later runs skip those articles. Skipped articles don't count toward maxItems. Dedupe is off by default because one-off pulls are the most common use.

Is scraping TechCrunch allowed? The actor reads publicly available pages and API endpoints. TechCrunch's terms restrict republishing its content, so personal, research and internal-analysis use is the typical case. You are responsible for how you use the data. The actor doesn't store or republish anything outside your own Apify storage.

Something broke or a field is missing? Open an issue on the actor's Issues tab and include the run ID.