Cloudflare

What is Cloudflare?

Cloudflare + AI lets you run fast, low-latency AI inference using pre-trained machine learning models directly on Cloudflare Workers, deployed across Cloudflare’s global network.

The platform provides full-stack AI building blocks: serverless AI on GPUs, a selection of popular models, and the ability to run models from Workers, Pages, or anywhere via its REST API. AI Gateway improves reliability and scalability with features like caching, rate limiting, and analytics. Vectorize, a globally distributed vector database, stores embeddings for efficient searches over your data.

Cloudflare + AI emphasizes simplicity and quick deployment, offering templates from a curated list of pre-built models. Supported tasks include image classification, sentiment analysis, speech recognition, text generation, and translation. You can run inference on Pages, popular frameworks, or any API-based setup with minimal code.

Companies including Meta, Nvidia, Microsoft, Hugging Face, and Databricks use Cloudflare + AI. The platform aims to help you build AI infrastructure that is dependable, secure, and cost-effective, without surprise bills. R2 provides affordable storage for training models and AI assets, making it cost-effective to set up multi-cloud environments for training large language models.

Who created Cloudflare?

Cloudflare + AI was developed by Cloudflare. It enables fast, low-latency inference on pre-trained machine learning models through Cloudflare Workers, combined with full-stack AI building blocks such as serverless AI on GPUs. Its goal is to support the creation and deployment of ambitious AI applications across Cloudflare’s global network.

What is Cloudflare used for?

• Running fast, low-latency inference on pre-trained machine learning models
• Building and deploying AI applications across Cloudflare’s global network
• Using serverless AI on GPUs and a selection of popular models
• Improving reliability and scalability with AI Gateway caching, rate limiting, and analytics
• Generating and storing embeddings in Vectorize for efficient searches
• Deploying quickly with templates for common AI tasks
• Storing training models and AI assets affordably with R2