axintai is on PyPI: App Intents from Python

Today we published axintai to PyPI. Python developers can now define App Intents using the same declarative API as the TypeScript SDK — and compile to identical Swift output.

Install

bash

pip install axintai

Define an intent

python

from axintai import define_intent, param

send_message = define_intent(
    name="SendMessage",
    title="Send Message",
    description="Sends a message to a contact",
    params={
        "recipient": param.string("Who to message"),
        "body": param.string("Message content"),
    },
)

Compile

bash

axintai compile intent.py --out Generated/

The output is the same production-quality Swift you get from the TypeScript SDK: a proper AppIntent struct, Info.plist fragment, and entitlements XML.

Why Python?

AI agents are increasingly Python-first. LangChain, CrewAI, AutoGen — the agent orchestration ecosystem lives in Python. With axintai, a Python agent can programmatically generate App Intents and compile them to Swift without ever touching Node.js.

The Python SDK produces the same language-agnostic IR as the TypeScript compiler. Same validator. Same generator. Same Swift output. Two author languages, one compiler pipeline.

What's next

The Python SDK is at v0.1.0 — it covers the core define_intent() and param.* API. Entity support and the MCP bridge are coming in v0.2.0.

Try it: pip install axintai and let us know what you build.