
Cloudflare's network now reaches more than 275 cities across over 100 countries, serving millions of customers and carrying more than 20 percent of the web. As the product set has expanded, the company has built separate abuse policies for different features, on the reasoning that a broad platform needs distinct rules for distinct services. The updated approach lives at the Abuse Hub, and the section below walks through how the policies are organized.
Guiding principles
The policies draw on input from human rights experts, activists, academics and regulators, and rest on a few commitments: rules should be specific to the service involved, so that actions address the actual harm with minimal unintended effects; anyone filing a complaint should have an abuse process that reaches the people able to resolve it narrowly, anonymously if needed; and the company should be transparent about both its policies and the actions it takes.
Three product categories, three approaches
Cloudflare's offerings fall into three rough buckets:
- Hosting products — Cloudflare Pages, Cloudflare Stream, Workers KV, Custom Error Pages.
- Security services — DDoS Mitigation, Web Application Firewall, Cloudflare Access, Rate Limiting.
- Core Internet technology services — Authoritative DNS, Recursive DNS/1.1.1.1, WARP.

Policy differs across each bucket, as described below.
Hosting: content rules where Cloudflare is the host
Hosting products are those where Cloudflare itself hosts the content, unlike security or temporary caching services where the content lives elsewhere. Abuse complaints and actions involving hosting are currently relatively rare, since most customers do not yet use these products. Because disabling access to hosted content effectively takes it offline, these products fall under the Acceptable Hosting Policy. Under it, Cloudflare may remove or disable access to content it believes:
- Contains, displays, distributes, or encourages the creation of child sexual abuse material, or otherwise exploits or promotes the exploitation of minors.
- Infringes on intellectual property rights.
- Has been determined by appropriate legal process to be defamatory or libelous.
- Engages in the unlawful distribution of controlled substances.
- Facilitates human trafficking or prostitution in violation of the law.
- Contains, installs, or disseminates any active malware, or uses the platform for exploit delivery (such as part of a command and control system).
- Is otherwise illegal, harmful, or violates the rights of others, including content that discloses sensitive personal information, incites or exploits violence against people or animals, or seeks to defraud the public.
Enforcement leaves room for discretion and aims to apply restrictions as narrowly as possible. A shopping cart platform with millions of customers, for example, would not lose Workers KV wholesale because one of its customers violated the policy. The underlying principle: organizations closest to the content are best placed to judge when it is abusive, and overbroad takedowns carry significant unintended consequences for access to content online.
Security services: widely available, rarely terminated
The overwhelming majority of Cloudflare's customers use only security services, and most pay nothing. The company's early decision to make security tools as widely available as possible — many free or at minimal cost — reflects both a practical goal of blunting cyberattacks and a view that attacks are not an acceptable mechanism for addressing problematic content. Attacks, in this framing, belong in the dustbin of history.
Wide availability forced careful thought about whether and when to terminate access, including the precedent any takedown sets. The conclusion, reached through conversations inside and outside the company, is that voluntarily cutting off services that protect against cyberattack is the wrong approach.
The abuse-of-power argument
The alternative — terminating services for reprehensible content so that others can knock it offline — is compared to a fire department declining to respond to a house whose owners lack moral character. That is a dangerous precedent, and over time it is most likely to harm vulnerable and marginalized communities disproportionately. With more than 20 percent of the web relying on these services, any policy also sets precedent for the Internet as a whole: terminating protection on the basis of personal moral judgment may be popular, but it makes it harder to defend content supporting oppressed and marginalized voices against attack.
What two terminations taught
Thousands of times per day, Cloudflare receives calls to terminate security services over content someone finds offensive; most never make news, and most do not conflict with the company's moral views. Twice, the company did terminate services over content it found reprehensible: the neo-Nazi troll site The Daily Stormer in 2017 and the conspiracy theory forum 8chan in 2019.
After both, Cloudflare saw a dramatic increase in authoritarian regimes attempting to have it terminate security services for human rights organizations, often citing the company's own justification language. Subsequent discussions with policymakers worldwide led to the conclusion that the power to terminate security services was not one Cloudflare should hold — not because the content was acceptable, but because security services most closely resemble Internet utilities. As with a telephone company that does not cut your line over awful or bigoted speech, turning off security services over despicable publications is the wrong policy. Having done it before does not make it right, or mean it will happen again.

Cloudflare still works to protect those targeted online. Project Galileo provides free cybersecurity services to nonprofits and advocacy groups, and the Athenian Project helps protect election systems in the United States and abroad. Election infrastructure must be fundamentally trustworthy and neutral, and content judgments that could be read as political would undermine that protection.
Legal limits and regulatory guidance
Content regulation laws passed around the world over the past five years largely distinguish services that host content from those providing security and conduit services, exempting the latter from acting as moderator without legal process. Cloudflare's policies follow that guidance. It blocks sanctioned organizations and individuals from using security services, and it terminates security services for content illegal in the United States, where Cloudflare is headquartered — including CSAM and content subject to the Fight Online Sex Trafficking Act (FOSTA). Otherwise, the position is that everyone should be free of cyberattacks, even where the company fundamentally disagrees with the content.
Cloudflare follows legal process controlling security services. Where a court in a country prohibits access to content, the company generally restricts access in that country per the order, recognizing that illegality in one jurisdiction does not make content illegal in another and tailoring restrictions narrowly to the relevant court's or authority's jurisdiction.
Transparency accompanies these restrictions: where possible, Cloudflare links to the specific legal order requiring the restriction. That matters for participation in the legal and legislative process. The company is sharply critical of ISPs that comply with court orders by invisibly blackholing content, leaving users no idea what legal regime prohibits it. Speech can be curtailed by law, but the Rule of Law requires whoever curtails it to be transparent about why.
Core Internet technology: a higher bar
For services like Authoritative DNS, Recursive DNS/1.1.1.1 and WARP, Cloudflare applies a higher bar than it does to legal orders affecting security and conduit services. Restrictions on these services are global in nature — they cannot easily be confined to one jurisdiction, so the most restrictive law ends up applying everywhere. Cloudflare has generally challenged or appealed orders attempting to restrict these services, even where a ruling applies only to free customers, and has suggested more tailored alternatives to regulators and courts.
Such cases are becoming more common, with copyright holders in particular seeking rulings in one jurisdiction that apply worldwide to terminate core services and wipe content offline — a precedent that would hand control over what content is allowed online to whichever jurisdiction is willing to be most restrictive. The company has so far largely succeeded in getting these cases overturned. But each exercise of discretion across security or core services weakens the argument in future cases.
Free and paid customers
Most customers use free services, and most abuse-process concerns involve free customers, but moderation policies do not differ based on whether a customer pays. Where a paying customer's values are diametrically opposed to the company's, Cloudflare takes further steps: it aims not to profit from the customer and to direct proceeds toward opposing causes. When a site opposed to LGBTQ+ rights signed up for paid DDoS mitigation, the company worked with its Proudflare employee resource group to identify an LGBTQ+ rights organization and donate 100 percent of the fees for its services. These efforts are not disclosed for marketing purposes.
Legal process, not popular opinion
Cloudflare believes it has an obligation to restrict content it hosts itself, but not the political legitimacy to determine generally what is online by restricting security or core Internet services. Harmful content, on this view, is properly restricted through legislation. An Internet where cyberattacks silence what is online is broken regardless of sympathy for the ends, so decisions about terminating security or core services are guided by legal process rather than popular opinion.
The company describes itself as not free speech absolutist but as believing in the Rule of Law: jurisdictions determine what content is allowed based on their own norms and laws. In assessing its obligations, Cloudflare looks to whether those laws are limited to the jurisdiction and consistent with its obligations under the United Nations Guiding Principles on Business and Human Rights.

Many injustices remain, and much online content is reprehensible. Cloudflare can solve some but not all of them; in improving the security and functioning of the Internet, the priority is to avoid causing it long-term harm. The company says it will keep the conversation going, cooperate with legitimate law enforcement investigating crimes, donate funds and services to equality and human rights causes, and take part in policy making worldwide to help preserve a free and open Internet.



