Ax LLM Quick Reference For Python

Use when writing Python code with axllm for using the generated Ax package, factory functions, package docs, examples, and API reference.

Install

Install only this skill for Python:

npx skills add https://ax-llm.github.io/ax/python/ --skill 'ax-python-llm'

Published skill file: ax-python-llm/SKILL.md.

Source

• Source: packages/python/skills/ax-python-llm/SKILL.md
• Version: 22.0.3

Skill Instructions

This skill helps an agent write Python code with the generated Ax package axllm. Use the generated package API, examples, and manifests; do not import TypeScript-only APIs unless you are editing the TypeScript package.

When To Use

• Start a generated-language Ax program from package docs or examples.
• Translate the Ax mental model into the target package without TypeScript-only imports.
• Choose the native package entrypoints for signatures, providers, generators, agents, flows, and optimizers.

Package Facts

• Language: Python.
• Package: axllm.
• Package API docs: API.md and axir-api.json.
• Capability manifest: axir-capabilities.json.
• Runnable examples: examples/.
• Real network support: yes.
• Scripted no-key transport support: yes.
• Runtime profiles: javascript-quickjs, python-pyodide.

Core Pattern

import os
from axllm import ai

llm = ai("openai", api_key=os.environ["OPENAI_API_KEY"])

Relevant API Surface

• Signatures: s, f, AxSignature
• AxGen: ax, AxGen
• AxAI: ai, OpenAICompatibleClient, OpenAIResponsesClient, GoogleGeminiClient, AnthropicClient, AxBalancer, MultiServiceRouter, ProviderRouter
• Agents And RLM: agent, AxAgent
• Flow: flow, AxFlow
• Optimizers: optimize, AxBootstrapFewShot, AxGEPA, OptimizerEngine, OptimizerEvaluator

Guardrails

• Start from package examples for exact native syntax before inventing a new call shape.
• Use provider-api examples only when the user explicitly has provider credentials available.
• Use no-key examples for deterministic local checks and provider request mapping.
• Treat AxIR as the source of generated package truth: if package docs disagree with source code, update the compiler and regenerate packages.
• Do not copy repo-maintainer skills from tools/*/skills/ into user packages.