Offline-First & Local-First Design

DevDiff is designed from the ground up to respect developer privacy and work reliably in offline environments (such as flights, remote areas, or secured company intranets).

Local Execution Guarantees

1. No External Telemetry: DevDiff does not send telemetry, usage statistics, or crash reports to any external servers.
2. Local AI Providers: By using WebGPU, WebAssembly, or a local Ollama instance, all LLM inference happens entirely on your local machine.
3. No Cloud Requirement: DevDiff operates with zero costs and zero dependencies on cloud APIs.

Config Configuration

To enforce local-first mode, ensure localOnly is enabled in your configuration:

export default {
  ai: {
    routing: {
      strategy: "priority",
      localOnly: true,
    },
  },
};

If a cloud provider is used in localOnly mode, the PrivacyEnforcer will automatically block the request or fallback to local inference.