Extract bank statements, invoices, and receipts with DocuClipper, then let your AI agent post them to QuickBooks. End-to-end MCP workflow from Claude Desktop, Cursor, or Continue.
Pair with a QuickBooks MCP server (third-party or your own) and your agent can chain extract → post.
{
"mcpServers": {
"docuclipper": {
"command": "npx",
"args": ["-y", "docuclipper-mcp"],
"env": {
"DOCUCLIPPER_PAT": "dcp_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
}
}
}User: Process the 12 monthly Chase statements in ~/2025-statements/ and post the transactions to QuickBooks Online (Acme Corp realm). Skip any statement that doesn't reconcile.
Claude (chains docuclipper.convert_bank_statement -> qbo MCP tool):
- 12 statements processed; 12/12 reconciled
- 1,247 transactions queued for QuickBooks
- Posted to QuickBooks: 1,247 transactions
- Done in 4m 12sSame extraction engine as the DocuClipper REST API, with an AI agent in the loop.
DocuClipper's MCP server pulls structured data from PDF bank statements, invoices, and receipts. Once extracted, push to QuickBooks Online via the agent's QuickBooks tools or the DocuClipper UI.
Drop a folder of statements into Claude Desktop, ask the agent to extract them and post to QuickBooks. The agent chains MCP tools without manual switching.
Each statement returns a reconciled flag so the agent only posts files where transactions match the printed balance, exactly the same QA model used in production accounting workflows.
Any Model Context Protocol client that speaks stdio can call DocuClipper. Add docuclipper-mcp to your config and the tools appear next to your other MCP servers.
Generate a Personal Access Token in DocuClipper, drop it in DOCUCLIPPER_PAT. Your agent inherits your DocuClipper credit balance and document history.
command: npx, args: [-y, docuclipper-mcp]. Pair with your favorite QuickBooks MCP server and let the agent move data between them.