---
id: toolbox-core
version: "1.3.0"
license: unclear
license_treatment: permissive
maintenance: active
---

# toolbox-core — Python Base SDK for interacting with the Toolbox service

License: permissive · Maintenance: active · Downloads: 437.6K/mo

## What it is and what it does

Toolbox Core is a Python SDK that bridges your GenAI application to a running MCP Toolbox service, allowing you to load tool definitions from the service and invoke them as standard Python functions. It provides both async-first and synchronous interfaces, with built-in support for parameter binding, authentication (including google-auth integration), and OpenTelemetry observability.

The package depends on pydantic for data validation, aiohttp and requests for HTTP communication, and the deprecated library for backward compatibility. It targets modern Python versions (3.10 through 3.14) and is designed to be framework-agnostic, working standalone or integrated into frameworks like Langraph.

Use it for:

- Load weather, calendar, or search tools from a Toolbox service and call them directly from a GenAI agent or chatbot.
- Integrate external tool definitions into a Langraph workflow without reimplementing tool interfaces.
- Build a multi-tenant GenAI application where tools are centrally managed in a Toolbox service and accessed by multiple clients.
- Add observability to tool invocations via OpenTelemetry without modifying tool definitions.
- Authenticate tool calls to a Toolbox service using google-auth or custom credentials.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

A Python SDK for connecting to and invoking tools from a running MCP Toolbox service, loading tool definitions as callable Python functions for use in GenAI applications.

Yes. The package is actively maintained, has low install friction, carries a permissive license, and solves a specific integration problem for GenAI applications that need to consume tools from a Toolbox service. No known vulnerabilities. Install if you are building against an MCP Toolbox deployment.

## Install

pip install toolbox-core
uv add toolbox-core
poetry add toolbox-core

## Installing toolbox-core

Before you install: Low friction installation with a pure-Python wheel. Active maintenance with a recent release (17 days old) and an actively developed repository.

License in practice: Licensed under Apache License 2.0 (permissive), allowing commercial and private use with minimal restrictions.

Quickstart:

pip install toolbox-core

import asyncio
from toolbox_core import ToolboxClient

async def main():
    async with ToolboxClient("http://127.0.0.1:5000") as toolbox:
        tool = await toolbox.load_tool("get_weather")
        result = await tool(location="London")
        print(result)

asyncio.run(main())

Requires a running MCP Toolbox service accessible at the specified URL; Python 3.10 or later.

Verify before relying:

- Whether synchronous wrapper utilities are built-in or require manual wrapping for non-async contexts.
- Support scope for authentication methods beyond google-auth (e.g., API keys, custom tokens).
- Whether tool parameter binding supports complex nested structures or only scalar types.

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 437.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

mcp toolbox python client, load remote tools as functions, genai tool integration sdk, toolbox service python sdk, async tool invocation framework, genai-integration, async-first, tool-orchestration

[View on SkillFed](https://skillfed.io/packages/toolbox-core) · [View on PyPI](https://pypi.org/project/toolbox-core/)