Crunchbase's terms forbid any use that crawls, scrapes or spiders its pages or data, by hand or by machine. For data in bulk, Crunchbase sells CSV exports on its paid plans, an API with unlimited exports, and data licenses. Two datasets are free under a Creative Commons license. A proxy does not change any of this.
What Crunchbase lets you export
Crunchbase's help center sets out the export limits by plan:
| plan | exports, in Crunchbase's words |
|---|---|
| Pro | In one line "up to 1k rows of data to a CSV file per month", and in the next "2k exports per month". |
| Business | Customers "can export up to 5k rows of data from their searches and lists to a CSV file per month". |
| API | "Unlimited exports", through Crunchbase's sales team. |
| free trial | No exports: "exporting capabilities are not available during the free 7-day subscription trial period". |
The article on the Pro trial also says "up to 2k exports". The 1k figure appears once in the articles we read, so ask Crunchbase before you rely on either number. An export is a picture of the search or list at the time you export it.
API customers can also take a CSV export without writing code. Crunchbase describes it as "a daily snapshot of the contents of each API endpoint collection", with separate files for companies, people, funding rounds and more. It is updated every 24 hours, and your API package decides which files you can download. A help article edited in February 2026 calls this a legacy method. The daily snapshot now comes as NDJSON, which the article recommends, or as a flat CSV.
The free datasets
Crunchbase's terms license two datasets under the Creative Commons Attribution License: the Crunchbase 2013 Snapshot and Crunchbase Basic. If you keep to the terms, you may use both under that license. The Snapshot must be credited to "CrunchBase 2013 Snapshot" with the notice "© 2013". Crunchbase says the credit line avoids confusion with its current dataset. The terms give no download address, and they name policy@crunchbase.com for questions.
What Crunchbase's terms say
The terms of service were last updated on September 30, 2026. They list uses that are not allowed, and these four decide this question:
| rule | what the terms forbid |
|---|---|
| scraping | Any use that "crawls," "scrapes," or "spiders" any page, data or portion of the service, by manual or automated means. |
| export limits | Getting around features, limits or restrictions, including any access, download or export of content with automated tools. |
| copying | Copying or storing a significant part of the content, including into a third-party tool such as ChatGPT, except as the Service expressly permits it or through Crunchbase's MCP Server. |
| model training | Using the content to train models, including generative AI. |
- Scrapingany use that crawls, scrapes or spiders a page or its data
- Getting around export limitsor any other feature, limit or restriction
- Copying a significant part of the contentincluding into a third-party product
- Training models on the contentgenerative AI included
A breach "is grounds for account suspension or termination of your right to use or access the Service". For any other use of the content, the terms send you to Crunchbase's products page, where its API and data licenses are listed.
What robots.txt says
Crunchbase's robots.txt sorts crawlers into groups. For every crawler it does not name, it closes search and discovery pages, account pages and the /v4 data addresses. A group of AI answer agents may crawl every path, under a comment that says "They cite us". AI training and bulk crawlers are blocked, under a heading that says they need a "license to access".
What crunchbase.com did when our server asked
We asked Crunchbase for pages from our server in the United States on October 3, 2026. The home page and robots.txt answered normally. The terms page on the main site answered HTTP 403 from Cloudflare, and the developer docs answered with a Cloudflare challenge. We read the terms on Crunchbase's about site, and the help articles through the help center's public interface for articles.
Where a proxy fits
For collecting Crunchbase data, a proxy does not fit. The terms forbid crawling and scraping by any means, and Crunchbase sells exports, an API and data licenses for that job. The two free datasets need no proxy either.
For other work where a proxy is allowed, what a residential proxy is explains the type. How to check if a proxy is working shows how to test one. Residential traffic without city targeting starts at $0.44/GB, and every plan is on our pricing page.
The limits of this page
We read Crunchbase's terms, help articles and robots.txt as text, from our server on October 3, 2026. We did not sign in, export anything, or call the API. The API's price and its rate limits are not stated on the pages we read. Crunchbase's own articles give different Pro figures. This page is due for a check by January 4, 2027.
Sources
Every page below was read on October 3, 2026, from our server in the United States.
- Crunchbase, Terms of Service, last updated September 30, 2026.
- Crunchbase Support, Export to CSV, "Export Searches or Lists to CSV", "What is the CSV Export?", "CSV Export FAQ", "Try Crunchbase Pro for Free" and "Accessing via CSV" (edited February 26, 2026).
- Crunchbase, Crunchbase API product page, and the robots.txt file of crunchbase.com.
- Our requests to crunchbase.com from our server in the United States, October 3, 2026.
