Extract Transactions from Images via Telegram & unli.dev Vision API

This n8n workflow turns images of receipts or transactions shared via Telegram into structured, Markdown-formatted text using the unli.dev Vision API. When a photo is sent to your Telegram bot, the workflow downloads it, converts it to base64, sends it to the API for intelligent analysis with a customizable prompt (default: extract transactions in Markdown), and replies directly in the chat with the extracted details. Optionally, it can save the data to Google Sheets, Airtable, or a database for further processing.

Key benefits include saving up to 90% of manual entry time, eliminating errors from blurry screenshots, and providing under-15-second turnaround per image. It offers a competitive edge by delivering instant structured feedback to staff or customers, making it ideal for streamlining data intake without typing.

Perfect for small business owners capturing payments, accountants handling client receipts, and ops teams converting screenshots to actionable records. Setup is quick: import the JSON, add Telegram bot token and unli.dev API key, customize the prompt, and test with a real photo. Requires n8n instance; optional storage integrations enhance persistence.