OCR Receipts to Notion: Structured Metadata Generation

Name: OCR Receipts to Notion: Structured Metadata Generation
Rating: 5 (5 reviews)
Author: Free N8N

Beginner

0 nodes connected

detail.loadingPreview

Free N8N Temples

65 views

0 downloads

Data ManagementAIAutomationLangchainMetadataMultilingualNotionOCRSupabase

Automates the extraction and structuring of receipt data into Notion, generating metadata in English and Chinese.

🚀Ready to Deploy This Workflow?

⚡Deploy on Zeabur 🎁Get $200 Credit on DigitalOcean

About This Workflow

This workflow processes OCR'd receipt data, structures it, and stores it in Notion. It leverages AI to generate metadata in both English and Chinese, enabling multilingual searching and organization within Notion. The process involves receiving data via a webhook, splitting and embedding it, and then using an AI agent to process and store the structured information.

Key Features

Webhook Trigger: Accepts OCR receipt data via a POST request.
Text Splitting & Embedding: Prepares text data for AI processing using Langchain's Text Splitter and OpenAI Embeddings.
Vector Storage: Utilizes Supabase as a vector store to efficiently store and retrieve embedded data.
AI Agent for Structuring: Employs an Anthropic Chat Model and a RAG Agent to extract, structure, and generate metadata from the receipt data.
Multilingual Metadata: Generates metadata in both English and Chinese.
Notion Integration: Appends structured data and status logs to Google Sheets (acting as a proxy for Notion data ingestion in this example).
Error Alerting: Notifies via Slack in case of workflow errors.

How To Use

Trigger Workflow: Send OCR'd receipt data to the configured webhook URL (ocr-receipts-to-notion). The data should be in a format that can be processed by the Text Splitter node.
Data Processing: The workflow automatically splits the text, generates embeddings using OpenAI, and stores them in Supabase.
AI Structuring: An Anthropic Chat Model, augmented by the Supabase vector store (via Vector Tool and RAG Agent), processes the data. The RAG Agent is configured with a system message to guide its behavior.
Metadata Generation: The RAG Agent is responsible for extracting relevant information and generating structured metadata, including English and Chinese versions.
Data Persistence: The structured output from the RAG Agent is appended to a Google Sheet (referenced by SHEET_ID and Log sheet name). This sheet acts as the destination for the processed and enriched data.
Error Handling: If any part of the RAG Agent execution fails, a Slack alert will be sent to the #alerts channel.

Apps Used

Automation

Langchain

Metadata

Multilingual

Notion

OCR

Supabase

Workflow JSON

{
  "id": "22e56381-83e3-4fa7-8c07-67312bb3cca8",
  "name": "OCR Receipts to Notion: Structured Metadata Generation",
  "nodes": 0,
  "category": "Data Management",
  "status": "active",
  "version": "1.0.0"
}

Note: This is a sample preview. The full workflow JSON contains node configurations, credentials placeholders, and execution logic.