Log expenses by talking to your AI. PocketLedger parses natural-language spending, saves transactions, and returns instant reports through a scoped MCP API.
Stop wrestling with spreadsheets. PocketLedger learns your habits and makes tracking effortless.
Just say "spent $45 at Trader Joe's on groceries yesterday" and PocketLedger parses the amount, merchant hint, category, date, and timezone when possible.
MCP toolReceipt and statement import are planned. Today, PocketLedger focuses on manual, API, widget, and natural-language transaction entry.
Coming soonTransactions keep their original currency, and reports group mixed currencies without pretending a live conversion happened.
Every create, update, and delete is reversible with full audit trail and before/after snapshots.
Spending breakdowns by category, merchant, day, week, or month with chart-ready structured data.
Use PocketLedger from Claude, ChatGPT, or any MCP-compatible assistant with OAuth2 scopes.
MCP nativeSee what happens from the moment you tell your AI about an expense to the database row.
Write a natural sentence with amount, merchant, category, and date.
The tool parses the text and extracts structured fields.
A transaction row is created with audit trail, duplicate checks, idempotency, and undo support.
Ask for totals, reports, category breakdowns, merchants, and trends.
Expose PocketLedger to any AI assistant via Model Context Protocol with fine-grained OAuth2 scopes.
"Logging expenses from chat is the first finance workflow that feels faster than opening a spreadsheet."
"The MCP tools make it easy to add a transaction, search history, and ask for summaries without leaving the assistant."
"Undo and soft-delete make the write tools feel safer when I am experimenting with assistant-driven workflows."
Connect your AI assistant in minutes and start tracking expenses in natural language.