Cloudflare selected 30 nonprofit organizations to participate in the first cohort of its program dedicated to civil society organizations, providing participants with credits worth a total of up to $7.5 million to use its development services. The company says these initiatives aim to help organizations build AI- and automation-based tools in the fields of human rights, humanitarian aid, mental health, education, and civic engagement.
The program comes as part of the expansion of Cloudflare for Startups to include nonprofits and public-interest organizations, with up to $250,000 in credits available per organization. Cloudflare also announced that applications are open for the program’s second cohort.
From Protecting Websites to Building Tools
The initiative is connected to Project Galileo, which provides free cybersecurity services to civil society organizations and protects more than 3,400 domains in over 120 countries. Cloudflare says the needs of these organizations have evolved from securing their existing applications to building new applications tailored to their workflows.
The tools rely on Cloudflare’s development services, including Workers AI, AI Gateway, and lightweight serverless infrastructure. According to the company, these services enable applications to run while scaling on demand, reduce the need for dedicated GPU infrastructure, and provide request-rate limits and monitoring and routing for AI model usage to reduce unnecessary calls.
However, the source explains that AI adoption still faces obstacles within the civil society sector. A CIVICUS survey indicated that more than half of respondents considered privacy concerns a barrier to use, while 48% said financial constraints limit adoption. These organizations handle highly sensitive data, such as activists’ locations and the identities of anonymous sources, making leaks or unauthorized access a problem that goes beyond operating costs.
Examples of Tools Developed by the Organizations
LebTown, an independent nonprofit newsroom in Lebanon County, Pennsylvania, developed an editorial management system that tracks stories from the proposal stage through publication and monitors reach and the impact of community work. It also built tools to convert property-transfer records from PDF files into structured data, as well as a copyediting agent that integrates with WordPress.
In India, Kaya Guides created a mental health application on WhatsApp that connects people with depression to trained counselors. By September 2026, the application had supported 10,439 people, including 2,325 currently registered users, and was processing around 500,000 messages per month. The organization also developed a system that listens to counseling calls and provides counselors with immediate feedback.
The Snorkelling Society, a British nonprofit organization, uses Cloudflare to secure the SnorkelMap application, which enables the global community of snorkeling enthusiasts to discover locations and add experiences and photos.
Automation in Sensitive Work
Cloudflare also worked with Freedom House to build a dashboard that captures and organizes public reports and identifies potential cases of transnational repression, configuring the tool according to the organization’s methodology and involving it in every stage of development. It is also working with the Global Network Initiative on a dashboard that monitors regulatory proposals and legal developments related to freedom of expression and privacy, and tracks advocacy opportunities and deadlines.
The partnership with Article One includes a human rights risk assessment tool that processes documents such as factory audits and country-context reports. Cloudflare emphasizes that AI helps organize information and suggest drafts, while final judgments remain in the hands of Article One staff.
What Matters in Practice?
The initiative offers a model for using AI in organizations that often lack engineering resources but handle data that cannot tolerate uncontrolled processing. The core value here is not merely the provision of cloud credits, but the combination of funding, development services, engineering support, and security and privacy controls.
Nevertheless, the examples do not show that automation eliminates the need for human review. Rather, they explicitly point to the need to identify tasks suitable for models, minimize data, prevent leakage and hallucinations, and keep humans involved in decisions with rights-related consequences. The initiative therefore appears important as a practical experiment in building AI tools for highly sensitive environments, rather than as a general solution to the problems of nonprofit organizations.