OpenAPITools SDK

What is OpenAPITools SDK?

OpenAPITools SDK is a Python package that enables developers to manage and execute tools across multiple AI API providers. It provides a unified interface for working with tools in Anthropic's Claude, OpenAI's GPT models, and LangChain frameworks.

How to use OpenAPITools SDK?

Install with pip install reacter-openapitools requests (plus langchain and langchain-core for LangChain). You need Python 3.8+, an API key from OpenAPITools, and access keys for at least one supported AI provider. Use either Local Mode (ToolsAdapter(folder_path="/path/to/tools")) or API Mode (ToolsAdapter(api_key="your_api_key")).

Key features of OpenAPITools SDK

• Create tools as Python or Bash scripts with standardized I/O
• Access tools through a single, consistent SDK
• Integrate tools with Claude, GPT, and LangChain models
• Build interactive chatbots that use tools for complex tasks
• Execute all tools locally – no code is sent externally

Use cases of OpenAPITools SDK

• Build AI assistants that can invoke custom tools to solve user requests
• Unify tool access across Anthropic, OpenAI, and LangChain with one SDK
• Run Python or Bash scripts as tools with minimal overhead
• Create interactive chatbots that combine reasoning with local tool execution

FAQ from OpenAPITools SDK

Which AI providers are supported?

Anthropic’s Claude, OpenAI’s GPT models, and LangChain frameworks are supported out of the box.

What types of tools can I create?

You can create Python tools (executed in-process via exec()) and Bash tools (run as subprocesses). Python tools receive arguments via an input_json dictionary; Bash tools get JSON on standard input.

What are the differences between Local Mode and API Mode?

Local Mode runs tools entirely from a folder on your machine and is preferred. API Mode uses an API key to communicate with OpenAPITools servers and applies rate limits.

How is security handled?

All tool execution happens locally within your environment. No code is sent to external servers. Environment variables can be securely passed to tools.

What are the runtime dependencies?

Python 3.8 or later, the reacter-openapitools and requests packages, and optionally langchain and langchain-core when using the LangChain adapter. You also need API keys for OpenAPITools and at least one AI provider.