Expense Budget Tracker MCP server
{
"mcpServers": {
"expense-budget-tracker": {
"type": "http",
"url": "https://mcp.expense-budget-tracker.com/mcp"
}
}
}{
"mcpServers": {
"expense-budget-tracker": {
"type": "http",
"url": "https://mcp.expense-budget-tracker.com/mcp"
}
}
}{
"mcpServers": {
"expense-budget-tracker": {
"type": "http",
"url": "https://mcp.expense-budget-tracker.com/mcp"
}
}
}{
"servers": {
"expense-budget-tracker": {
"type": "http",
"url": "https://mcp.expense-budget-tracker.com/mcp"
}
}
}{
"mcpServers": {
"expense-budget-tracker": {
"type": "http",
"url": "https://mcp.expense-budget-tracker.com/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.6.0
What it does
An MCP server that exposes expense tracking and budget management data from a self-hosted Postgres database. Developers can query transactions, budgets, and balances through restricted SQL, or insert and update financial records with write access. Authentication uses OAuth for MCP connections or API keys for direct agent access.
- Query transaction history and account balances across multiple currencies to generate spending reports or analyze financial trends.
- Allow AI agents to insert expense records parsed from bank statements, screenshots, or CSV files into the expense tracker.
- Build custom financial dashboards or analysis tools by querying the underlying ledger schema with read-only SQL access.
Tools this server exposes
| Tool | Does |
|---|---|
| list_workspaces | List workspaces available to the signed-in user. |
| get_schema | Inspect the allowed relations, columns, constraints, and limits of the expense tracker database. |
| sql_query | Run a restricted SELECT or WITH...SELECT statement to read financial data. |
| sql_execute | Run an approved INSERT, UPDATE, or DELETE statement to modify expense records. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.