Summary
- Build AI applications in Go with generation, structured output, streaming, tool calling, and flows across model providers.
- Core features include text and data generation, prompt templates, tool definition with interrupts, and streaming support across Google AI, Vertex AI, Anthropic, OpenAI, and Ollama Define flows for tracing, observability, and HTTP deployment via genkit.Handler ; test flows directly from the CLI or Developer UI Use middleware for cross-cutting concerns: built-in Retry , Fallback , ToolApproval , Filesystem , and Skills compose via ai.WithUse Pass the *Genkit instance explicitly to all SDK functions; use .prompt files with Handlebars templating for complex prompts and struct tags for structured output quality