Turn web content into useful client work.

No. The version that we provide is a self-hosted version, which is hosted on your server instead of Firecrawl.

Firecrawl Cloud offers a list of managed features in the form of Browser, Interact and hosted dashboards, which are managed on your behalf. The default self-hosted stack doesn't include those. You inherit the engine of crawling and scraping the core and the rest of it is your business to operate.

That's not necessarily a downside. Self-hosting implies that you have control over the version, resources and the location of data. It simply means that the conveniences you would enjoy with Cloud are not in the picture unless added on its own.

No, and it is better to say that simply.

The ability to read a page depends on three aspects: access controls set by the site itself, the structure of the page, and your deployment setup. Other websites simply prevent automated access. Certain load contents are in such a form that they require additional treatment. Others are okay, provided a particular setting or proxy exists.

We examine your real target sites when you are setting up and inform you what works. We cannot guarantee you access to blocked or restricted material; that is the choice of the site, and not a configuration issue.

Just use what you have a right to.

Public visibility in the web is not like freedom to reuse. The terms of service, copyright policies, and privacy requirements of the websites continue to apply to the content collected and they do not vanish because a tool has collected the content on your behalf.

Practically, that entails being conscious of what sources you are pointing it at, and what you do with the results. Internal research, briefs and enrichment are most of the time easy. Whether to republish the content of another person is a separate issue and one that is worth inquiring before you come to trust it.

That is, according to your agreement, which is where the detail resides.

There, infrastructure, server resources, access arrangements, updates, and backups and support are all defined. This may be run and supported by us, as a whole, installed, documented, and handed over depending on the arrangement.

One fact to be aware of: a separate installation can be used on shared hardware. That saves on expenses, yet it results in pooled resources with additional deployments. Your consent will inform you of what arrangement is applicable to you.

Where your order tells you so.

The cost of running Firecrawl alone is one thing. Anything that it is linked to is another. Each AI process, proxy service and any third-party tool may require its own account and have its own usage fees, charged by the provider.

Those, too, are volume-dependent costs. Minor research workflow scarcely amounts. Its workflow on hundreds of pages per day is another figure. We would prefer to create such an expectation prior to setup than following the initial invoice.

No, and it is worth correcting that supposition.

Firecrawl makes calls to the sites it visits; that is the entire objective. Whatever it gathers must have an external source of your server. In case the output is then forwarded to a related AI service or via a proxy, that data is being accessed by a third party as well.

Depending on your configuration: which services are attached, what each one has been allowed to accept, how long data is stored, and whether the setup is the one that ensures compliance with your requirements. Those get signed on your targeted deployment; not everybody can get the same answers.