Frequently Asked Questions (FAQ) about UBOS Asset Marketplace’s MCP Server

Q: What is an MCP Server? A: MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs. An MCP Server acts as a bridge, allowing AI models to access and interact with external data sources and tools.

Q: What is the Unstructured API MCP Server? A: It’s a specific MCP server implementation designed for interacting with the Unstructured API. It provides tools to list sources, destinations, workflows, and manage data processing jobs.

Q: What tools are available in the Unstructured API MCP Server? A: The server provides tools for listing, getting info, creating, updating, and deleting sources, destinations, and workflows. It also allows you to list, get info, and cancel jobs.

Q: What data sources and destinations does the MCP Server support? A: Currently, the server supports S3, Azure, Google Drive, OneDrive, Salesforce, and Sharepoint as sources. Destinations include S3, Weaviate, Pinecone, AstraDB, MongoDB, Neo4j, Databricks Volumes, and Databricks Volumes Delta Table. See full list of supported sources here and destinations here

Q: What is Firecrawl and how does it integrate with the MCP Server? A: Firecrawl is a web crawling API that allows you to retrieve HTML content or generate LLM-optimized text from websites. The MCP Server provides tools to invoke Firecrawl crawls and check their status.

Q: What credentials are required to use the connector tools? A: Each connector requires specific credentials defined in your .env file. Refer to the documentation for a list of required credentials for each connector.

Q: How do I install and configure the Unstructured API MCP Server? A: The installation process involves using Python 3.12+, installing the uns_mcp package, configuring Claude Desktop (if using), and setting the necessary environment variables, including your Unstructured API key.

Q: What is UBOS, and how does the MCP Server fit into the UBOS platform? A: UBOS is a full-stack AI Agent Development Platform. The MCP Server integrates into UBOS by providing a way to connect AI Agents with external data sources and tools, enabling more complex and data-driven AI applications.

Q: Can I use a subset of tools with my client? A: Yes, you can. However, if you’re using update_workflow, it must be loaded in the context with create_workflow because it provides details on how to create and configure custom nodes.

Q: Where can I find the changelog for updates and new features? A: All new features, fixes, and enhancements are documented in the CHANGELOG.md file.