What it does
A Telegram workflow that reads receipt photos with Tesseract OCR and processes text messages directly. A LLaMA model through OpenRouter categorizes each expense, such as Food & Beverages, Household or Transport, and replies with a structured summary. It uses a community node, so it needs a self-hosted n8n instance.
Use cases
- 01Saving an expense by sending only a receipt photo
- 02Sorting spending into categories such as household or transport
- 03Getting a clear summary back after each receipt
Connects
HTTP Request · Telegram · Code · Basic LLM Chain · Structured Output Parser · OpenRouter Chat Model