Examples index

This page indexes runnable programs under example/llm/ in the repository. For shorter task-oriented examples, use the LLM recipes.

Running an example

Each directory is a separate main package using the root module:

go run ./example/llm/basic

Live requests require the selected provider's API key. Check Protocol and provider status first.

Example programs

Example	Source directory	Demonstrates	Documentation
Basic	example/llm/basic	LookupModel, Complete, text results	Basic completion
Options	example/llm/options	System prompt, temperature, max tokens	Configuration
Streaming	example/llm/streaming	Text deltas and terminal events	Streaming responses
Reasoning	example/llm/reasoning	Reasoning effort and thinking events	Requesting reasoning
Tools	example/llm/tools	Typed tools and an application tool loop	Executing tool calls
Conversation	example/llm/conversation	Caller-owned multi-turn history	Messages and context
Model switch	example/llm/model_switch	Changing protocol and model with one history	Changing models in a conversation
Advanced	example/llm/advanced	Request hooks and low-level options	Configuration
Providers	example/llm/providers	Provider registration and overrides	Providers
Who am I	example/llm/whoami	Provider credential status	Provider status

Compile check

This command compiles every tracked LLM example without sending a network request:

go test ./example/llm/...

Examples cover llm responsibilities only. A production application still owns session storage, tool authorization, retry policy, and context compaction.