Unstructured

You can run the Unstructured API MCP Server to manage sources, destinations, and workflows for the Unstructured platform. This server exposes tools to list, create, update, and delete connectors, workflows, and jobs, enabling automated control of your data pipelines and processing tasks.

How to use

You will interact with the MCP server through an MCP client. Start the server using the recommended environment management flow, then connect your client to perform actions such as listing sources, creating connectors, configuring workflows, and running or monitoring jobs. Use the available tools to list resources, inspect details, and execute operations in a controlled sequence.

How to install

Prerequisites you need before starting:

• Python 3.12+
• uvx for environment management
• An API key from Unstructured to access the MCP server

Option A — Using SSE Server Protocol (recommended for debugging and local testing): use the following command to start the server in one terminal and then test with a local client in another terminal.

run python uns_mcp/server.py --host 127.0.0.1 --port 8080
# or
make sse-server

Option B — Using Stdio Server Protocol (Claude Desktop integration): configure your MCP client to run the stdio server with the following example. This runs the UV tool and points it to the local MCP repository containing the server code.

{
  "mcpServers": {
    "UNS_MCP": {
      "command": "ABSOLUTE/PATH/TO/.local/bin/uv",
      "args": [
        "--directory",
        "ABSOLUTE/PATH/TO/YOUR-UNS-MCP-REPO/uns_mcp",
        "run",
        "server.py"
      ]
    }
  }
}

If you prefer to run the local client for testing, you can execute the following command to connect to the SSE endpoint once the server is running.

uv run python minimal_client/client.py uns_mcp/server.py

Additional configuration and notes

The server interacts with a variety of connectors and workflows. To manage these, you will use a combination of commands to list, inspect, create, update, and delete resources. Be sure to set up the necessary credentials in your environment for the connectors you intend to create or modify.

Environment variables required for connector operations are listed below. You should define these in your environment or in a .env file that your MCP server reads.

ANTHROPIC_API_KEY
AWS_KEY
AWS_SECRET
WEAVIATE_CLOUD_API_KEY
FIRECRAWL_API_KEY
ASTRA_DB_APPLICATION_TOKEN
ASTRA_DB_API_ENDPOINT
AZURE_CONNECTION_STRING
AZURE_ACCOUNT_NAME
AZURE_ACCOUNT_KEY
AZURE_SAS_TOKEN
NEO4J_PASSWORD
MONGO_DB_CONNECTION_STRING
GOOGLEDRIVE_SERVICE_ACCOUNT_KEY
DATABRICKS_CLIENT_ID
DATABRICKS_CLIENT_SECRET
ONEDRIVE_CLIENT_ID
ONEDRIVE_CLIENT_CRED
ONEDRIVE_TENANT_ID
PINECONE_API_KEY
SALESFORCE_CONSUMER_KEY
SALESFORCE_PRIVATE_KEY
SHAREPOINT_CLIENT_ID
SHAREPOINT_CLIENT_CRED
SHAREPOINT_TENANT_ID
LOG_LEVEL
CONFIRM_TOOL_USE
DEBUG_API_REQUESTS

Note: A key is required to access and use the Unstructured features. Ensure you keep credentials secure and do not expose them in logs or shared environments.

## Available tools

### list\_sources

Lists available sources from the Unstructured API.

### get\_source\_info

Get detailed information about a specific source connector.

### create\_source\_connector

Create a source connector.

### update\_source\_connector

Update an existing source connector by params.

### delete\_source\_connector

Delete a source connector by source id.

### list\_destinations

Lists available destinations from the Unstructured API.

### get\_destination\_info

Get detailed info about a specific destination connector.

### create\_destination\_connector

Create a destination connector by params.

### update\_destination\_connector

Update an existing destination connector by destination id.

### delete\_destination\_connector

Delete a destination connector by destination id.

### list\_workflows

Lists workflows from the Unstructured API.

### get\_workflow\_info

Get detailed information about a specific workflow.

### create\_workflow

Create a new workflow with source, destination id, etc.

### run\_workflow

Run a specific workflow with workflow id.

### update\_workflow

Update an existing workflow by params.

### delete\_workflow

Delete a specific workflow by id.

### list\_jobs

Lists jobs for a specific workflow from the Unstructured API.

### get\_job\_info

Get detailed information about a specific job by job id.

### cancel\_job

Delete a specific job by id.