AI Integration

Create invoices from your AI assistant

Connect Claude Desktop, Cursor, or any MCP client to create, send, and track invoices using natural language. 18 tools, draft-first workflow, no subscription.

18 toolsOAuth 2.1$0.30 per invoiceFree drafts

Get Started View Documentation

Get started in minutes

Open Claude Desktop settings
Add the Billstride MCP server URL to your "mcpServers" config
Restart Claude Desktop. OAuth will prompt you to sign in.

json{
  "mcpServers": {
    "billstride": {
      "url": "https://billstride.com/api/mcp/sse"
    }
  }
}

18 tools for the full invoice lifecycle

From drafting to delivery tracking. More tools than any other invoicing MCP integration, with built-in analytics no competitor offers.

create_invoice

Create a new draft invoice with line items, tax rates, and currency

update_invoice

Update any field on an existing draft invoice

delete_invoice

Permanently delete a draft invoice

duplicate_invoice

Create a copy of an existing invoice with a new number

create_invoice_from_previous

Create a new invoice pre-filled from a previous one for the same client

get_invoice

read-only

Retrieve full details of a specific invoice by ID

get_invoice_pdf

read-only

Generate and return the PDF for an invoice

list_invoices

read-only

List all invoices with optional filters for status, client, and date range

suggest_invoice_number

read-only

Suggest the next invoice number based on existing patterns

How it works

Create a Draft

Ask your AI assistant to create an invoice. Drafts are free and include a DRAFT watermark on the PDF.

"Create an invoice for Acme Corp. 10 hours of consulting at €120/hour, 19% VAT"

Review & Finalize

Review the draft in Billstride or ask the AI to make changes. When ready, finalize the invoice. This charges $0.30 and removes the watermark.

"Change the due date to net 30 and add a 5% early payment discount"

Send & Track

Send the invoice by email directly from your AI assistant. Track delivery status, opens, and bounces.

"Send the invoice to billing@acme.com with a friendly reminder message"

Authentication

OAuth 2.1 is the recommended authentication method. MCP clients that support OAuth will automatically discover the authorization endpoints and prompt you to sign in.

Auto-discovery via .well-known/oauth-authorization-server
PKCE (Proof Key for Code Exchange) for security
One-time browser sign-in — tokens last 30 days
Automatic token refresh

http// Discovery endpoint
GET https://billstride.com/api/mcp/.well-known/oauth-authorization-server

// The MCP client handles the full OAuth flow automatically.
// You just need to approve access in your browser once.

Rate limits: 100 requests per minute per user. If you exceed the limit, requests return HTTP 429 with a Retry-After header.

Frequently asked questions

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude connect to external tools and services. It enables your AI to take actions, like creating invoices, on your behalf, with your explicit permission.

Which AI clients are supported?

Any MCP-compatible client works with Billstride. This includes Claude Desktop, Cursor, Windsurf, and other tools that support the MCP standard. New clients are adopting MCP regularly.

Does using MCP cost extra?

No. The MCP integration is free. You only pay the standard $0.30 per finalized invoice. Same price as using the web app. Drafts are always free.

Is my data safe?

Yes. All connections use OAuth 2.1 with PKCE for secure authentication. Your data is encrypted in transit and at rest. The AI assistant only has access to your Billstride account data, and you can revoke access at any time.

Can I manage the full invoice lifecycle from AI?

Absolutely. You can create, edit, duplicate, finalize, send, and track invoices entirely from your AI assistant. You can also check analytics, manage clients, and update your business settings.

What happens if I exceed the rate limit?

If you exceed 100 requests per minute, the API returns HTTP 429 (Too Many Requests) with a Retry-After header. Most AI clients handle this automatically by waiting and retrying.