functai

Write a Python function. A language model does the work. You measure how well.

functai turns a typed Python function into a call to a language model. The function's name, docstring and types say what you want; the answer comes back as the type you asked for. Then you run it on a whole table, find out how often it is right, and make it better.

from typing import Literal
from dpyr import col
import functai
from functai import ai

@ai
def team(message: str) -> Literal["shipping", "billing", "product", "account"]:
    """Which team should answer this customer message?"""

team("I was charged twice for order B-2210, please fix this.")    # 'billing'

tickets = functai.datasets.tickets()                              # 80 labelled support messages
tickets.mutate(team=team(col.message))                            # a new column, one call per message
functai.evaluate(team, tickets, expected="category")              # how often it's right, with a range

Installation

pip install "functai[data]"      # Python 3.11+

With an API key in your environment (OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, …) or a Claude, ChatGPT or Copilot subscription, there's nothing to set up: functai picks a small model you can use and tells you which. To choose: functai.configure(lm="claude-haiku-4-5").

Documentation

maximerivest.github.io/functai, with three ways in:

• I have a table of text: label, sort or score every row, check it, make it better.
• I have notes or documents: pull the facts out as columns, following your protocol.
• I have a prompt that works: send it exactly as it is, then add types, tables and tests.

The examples each solve one problem end to end.

Built on

lm15 (every provider, no SDKs), lmcc (how values are written into prompts and read back) and dpyr (tables).

Development

From git: pip install "functai @ git+https://github.com/MaximeRivest/functai#subdirectory=python".

This folder is the Python package; the repository around it holds the contract every language's FunctAI follows and the documentation site. In this folder: uv sync --all-groups, then uv run pytest (offline, a fake provider). The documentation runs against real models: .venv/bin/python tests/docs_live.py --render (costs cents).