Skip to content

Learn: For AI Agents

Axon was designed for a world where most code is written by machines. Its homoiconic S-expression syntax removes the ambiguity that trips up LLM code generation — no operator-precedence guessing, no bracket-matching gymnastics, no lifetime annotations to hallucinate. This track is the fastest way to understand why Axon is a good generation target and how to produce correct Axon.

  1. Quick Start Intro

    Compile your first S-expression in the browser, then install locally.

  2. Tour of Axon Intro

    Interactive, step-by-step walk through core syntax and semantics.

  3. Why S-Expressions? Intermediate

    The homoiconicity argument: why uniform syntax lowers generation error rates.

  4. Axon for Rust Developers — The Token Crisis Intermediate

    Why dense, human-ergonomic syntax is hard for LLMs, and how Axon avoids it.

  5. Best Practices & Idioms Intermediate

    The idiomatic patterns to bias generation toward — and anti-patterns to avoid.

  6. Architecture Overview Advanced

    How source maps directly to the AST — the property that makes tooling trivial.

Where to go next

Section titled “Where to go next”

Once you can generate correct Axon, the Backend Engineers track shows how to assemble those programs into a running service using the standard library and web framework. If you’re generating inference code specifically, the AI/ML Engineers track covers the memory and concurrency primitives that matter for throughput.