Examples¶
These examples grow from one local stream into distributed and tool-using agents. Each one focuses on an applied task and points out the part of A11 that makes it work.
| Example | What you will build |
|---|---|
| Stream data through an AsyncNode | Produce and consume an ordered asynchronous stream, including clean finalisation. |
| A WebSocket echo session | Connect a client and server session over WebSocket and stream an action result back. |
| Talk to a model | Call interact_with_llm, feed a conversation, and print model output as it arrives. |
| From a local run to a remote call | Move an action across a session without changing its schema or handler-facing API. |
| A toy agent with a tool | Register an A11 action as an LLM tool and let the model use its result. |
| Going distributed | Choose storage and transport boundaries for an agent that spans processes or machines. |
| Browser clients | Connect a TypeScript browser application to an A11 backend. |
| Local models on the web | Run a local model in a browser-facing application while retaining A11 action semantics. |
Runnable companion programs live in the repository’s
examples/ directory.