fabrix-ai 0.1.0

Graph-based agent framework powered by oauth-codex

Fabrix

Fabrix is a graph-based agent framework built on top of oauth-codex>=2.2.0.

Key Features

• Graph-based state transitions
• Structured state outputs via Pydantic
• Tool execution with strict, tool-specific argument schemas
• Tool execution with Pydantic parameter validation
• Async-only, streaming-first API

Installation

pip install fabrix

Quickstart

import asyncio

from pydantic import BaseModel

from fabrix import Agent
from fabrix.events import ReasoningEvent, TaskFinishedEvent, ToolEvent


class AddInput(BaseModel):
    a: int
    b: int


def add_numbers(payload: AddInput) -> int:
    return payload.a + payload.b


async def main() -> None:
    agent = Agent(
        instructions="You are a precise assistant.",
        tools=[add_numbers],
    )

    async for event in agent.run_task_stream("Use add_numbers to compute 3 + 9"):
        print(f"[step={event.step}] {event.event_type}")
        if isinstance(event, ReasoningEvent):
            print("reasoning:", event.reasoning)
        elif isinstance(event, ToolEvent):
            if event.phase == "start":
                print("tool call:", event.tool_name, event.arguments)
            elif event.result is not None:
                print("tool result:", event.result.model_dump())
        elif isinstance(event, TaskFinishedEvent):
            print("final:", event.final_output)


asyncio.run(main())

See examples/minimal/quickstart.py and examples/advanced/data_workflow.py for more.

Notes

• Structured state output is compiled to a Structured Outputs-compatible schema subset.
• Tool-call items use strict schemas (anyOf, no oneOf/const) for provider compatibility.
• Every object node is normalized to explicit properties + required consistency.
• Map-like dict[...] shapes may be narrowed to empty-object constraints for strict compatibility.
• Define tool parameters with explicit fields (Pydantic models recommended) for reliable validation.