What mark_notification_read does
Performs the mark notification read operation against the organization's data.
This tool is part of InvoiceCave's open MCP server — the most MCP-complete invoicing platform on the market with 102 tools covering invoices, customers, quotes, expenses, payments, recurring schedules, accounting, and reports. Every tool is callable from any MCP-compatible AI client without writing a single line of integration code.
How an AI uses mark_notification_read
Called by an MCP-compatible AI client (Claude Desktop, Claude Code, Cursor, Windsurf, Codex) when the user's prompt maps to this action.
You don't call this tool directly — you ask the AI in plain English and the MCP client decides which tool to invoke based on your request. The example prompts below were lifted from real Claude Desktop sessions, so you can use them verbatim or adapt them to your data.
Parameters
mark_notification_read accepts 1 parameter — 1 required, 0 optional. The AI fills these in from your natural-language prompt; you never write them by hand.
| Parameter | Type | Required | Description |
|---|---|---|---|
| id | string | required | Notification ID |
Request example
Under the hood, your AI client sends a JSON-RPC 2.0 tools/call request to InvoiceCave's MCP endpoint at https://www.invoicecave.com/api/mcp/mcp. Here's what a mark_notification_read call looks like:
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "mark_notification_read",
"arguments": {
"id": "id_a1b2c3d4"
}
}
}You won't write this yourself — it's shown so developers can see exactly what the MCP client generates. The arguments object maps 1:1 to the parameters above. Authentication is handled by the Authorization: Bearer header with your API key.
Example AI prompts
Type any of these into Claude Desktop, Claude Code, or Cursor — InvoiceCave's MCP server will route them to mark_notification_read automatically.
“Run the "mark notification read" action for me.”
Use it in 30 seconds
Related tools
More from the Notifications & Email Logs category.
FAQ — mark_notification_read
What is mark_notification_read?▾
mark_notification_read is one of 102 tools in InvoiceCave's MCP server. Performs the mark notification read operation against the organization's data. It's invoked automatically when an AI client like Claude or Cursor decides the user's request maps to this action.
Which AI clients can call mark_notification_read?▾
Any MCP-compatible client. We've tested it with Claude Desktop, Claude Code, Cursor, Windsurf, and Codex. The MCP protocol is open, so any AI tool that speaks MCP can call this endpoint with a valid InvoiceCave API key.
Is calling mark_notification_read secure?▾
Yes. Every call is authenticated by the API key you mint in the dashboard, scoped to a single organization, rate-limited per key, and recorded in the audit log with source='mcp'. Sensitive fields (SMTP passwords, encryption keys) are never returned to the AI client.
What parameters does mark_notification_read take?▾
mark_notification_read accepts 1 parameter — 1 required (id). The AI populates these from your plain-English prompt — see the parameter table and request example above.
Does mark_notification_read count against any quota?▾
MCP tool calls are unmetered on the free plan up to a generous monthly limit. Once you connect Stripe and start charging your customers, every call still counts against the same per-key rate limit but doesn't add to your bill.
Ready to wire this up?
Free plan — no credit card. 102 tools at your AI's disposal.