cloudflare

Use Workers for serverless functions at the edge, Pages for full-stack web apps with Git deploys, and Durable Objects for stateful coordination or real-time workloads. The Skill's decision trees route based on whether you need compute, hosting, or strongly-consistent state.

Use KV for high-read key-value data like config and sessions, D1 for relational SQLite queries, R2 for S3-compatible object storage, and Durable Object storage for strongly-consistent per-entity state. The Skill provides a storage decision tree comparing consistency, latency, and use cases.

Use the Workers AI binding with env.AI.run() and a model like @cf/meta/llama-3.1-8b-instruct for text generation or BGE models for embeddings. For RAG patterns, combine Workers AI embeddings with Vectorize as the vector database.

Yes, the Cloudflare Terraform provider v5 manages zones, DNS records, Workers, KV, R2, D1, and Pages as code. Note that v5 renamed many resources from v4, and you should never mix Terraform with wrangler.jsonc for the same resources.

Cloudflare limits, pricing, and API signatures change frequently, so the Skill explicitly instructs verifying against live docs at developers.cloudflare.com, the wrangler config schema, and workers-types packages. Reference files are starting points, not the source of truth.

Use Snippets for lightweight HTTP request/response modifications under 5ms CPU time and 32KB code size, included free on paid plans. Choose Workers when you need bindings like KV or D1, more than 5 subrequests, or complex application logic.