Python Toolset

Build connected agents with StackOne’s Python AI SDK & Toolset

Build AI agents that connect to Salesforce, Workday, BambooHR, Slack, Gmail, HubSpot, and 200+ other business systems. The Python SDK integrates natively with LangChain, CrewAI, OpenAI Functions, and LangGraph. It includes built-in support for LangSmith feedback streams and dynamic tool discovery.

Not sure if the AI Toolset (SDK) is right for you? Check out our Agent Protocols Guide to understand when to use AI Toolset (SDK) vs MCP vs A2A.

Why StackOne?

200+ Providers

Connect to Salesforce, Workday, Slack, Gmail, and more through one SDK

Multi-Tenant

Built-in customer isolation with account_id for each connection

AI-Native

Tools designed for LLM function calling and agent frameworks

Consistent Data

Normalized schemas: no provider-specific parsing needed

How It Works

Your agent connects to StackOne, which handles authentication and data normalization for 200+ SaaS providers like Salesforce, Workday, BambooHR, Slack, and more.

Component	Description
Your Agent	Built with LangChain, CrewAI, OpenAI, Vercel AI, or any framework
StackOne	AI Integration Gateway: handles auth, transforms data, manages connections
SaaS Providers	Salesforce, Workday, BambooHR, Greenhouse, Slack, Gmail, HubSpot, and 200+ more

Key Concepts

• Initialize a ToolSet: Create a StackOneToolSet instance to access tools for providers like BambooHR, Greenhouse, Slack, and more
• Filter Tools: Use provider glob patterns like "bamboohr_*", "slack_*" or the fetch_tools() method for provider/action/account filters
• Execute Tools: Call provider APIs using tool.call() to list employees from Workday, send Slack messages, or create Jira tickets
• Meta Tools: Discover and run new tools at runtime via natural language search
• Implicit Feedback: Stream tool quality signals to LangSmith with automatic or manual instrumentation
• Multi-Tenant: Each tool execution requires an account_id to identify which customer’s account to access

Python Framework Integrations

LangChain

Build sophisticated agents with LangChain

CrewAI

Multi-agent workflows and collaboration

OpenAI

Direct OpenAI function calling

Python Core Features

Tool Filtering

Glob patterns plus provider/action filters with fetch_tools()

Meta Tools

Dynamic tool discovery and execution

Installation

Set up the Python SDK

Basic Usage

Core concepts and fundamental usage patterns

Implicit Feedback

Automatically emit LangSmith events from every tool call

Next Steps

1. Learn about tool filtering
2. Explore Meta Tools
3. Configure implicit feedback streams

Resources

• GitHub Repository: Complete source code and examples
• Examples: LangChain, CrewAI, and meta tools examples
• Issues: Report bugs or request features
• PyPI Package: Package details and version history