Connect IBM wxflows to Data Sources

With watsonx.ai Flows Engine you can build tools out of any data source, and deploy them to an endpoint in the cloud. Tools built with watsonx.ai Flows Engine can be used in any agentic framework using the SDK for Python and JavaScript.

Use this when you need to turn an arbitrary data source into a callable tool and deploy it to a cloud endpoint that any agent framework can call, rather than writing and hosting a bespoke tool integration per framework. It is a tool-building and deployment platform, not a specific pre-built tool itself - beyond the small set of ready-made examples (exchange, wikipedia, google_books, math, weather), it expects you to build your own tools from your own data sources.

Ships a handful of ready-made tools - exchange, wikipedia, google_books, math, and weather - alongside documentation for building custom tools from any data source. Provides integrations with LangGraph, LangChain, watsonx.ai, and OpenAI for tool calling, and ships example applications: an end-to-end agent chat app, a text-to-SQL agent, a YouTube transcription agent, a math agent, a Model Context Protocol (MCP) example, RAG (retrieval-augmented generation), and webpage summarization.

Sign up for free access to watsonx.ai Flows Engine, then build tools using the Python or JavaScript SDK and deploy them to a cloud endpoint. Support and questions are handled via the project's Discord.

Teams building AI agents in LangGraph, LangChain, watsonx.ai, or OpenAI-based frameworks who need a way to turn arbitrary data sources into deployable, cross-framework tools without building separate integrations for each agent framework.