🤖 AI Invoice Extraction Agent

Turn your messy invoice images into structured data and organized files automatically! 🚀

📑 Table of Contents

• 🌟 Overview
• ✨ Key Features
• 🛠️ Architecture & Workflow
• 📋 Prerequisites
• ⚙️ Installation & Setup
  • 1. Telegram Bot Setup
  • 2. Google Cloud Setup
  • 3. OCR.space Setup
  • 4. OpenAI Setup
  • 5. n8n Configuration
• 🚀 Usage Guide
• 💾 Data Structure & Customization
• 🔧 Troubleshooting & FAQ
• 🤝 Contributing
• 📄 License

🌟 Overview

The AI Invoice Extraction Agent is a state-of-the-art, no-code automation workflow built on n8n. It acts as your personal finance assistant, living right inside your Telegram messenger.

Imagine this: You're at a lunch meeting, you pay the bill, and instead of stuffing the receipt in your wallet to be forgotten, you simply snap a photo and send it to your Telegram bot. Done.

This workflow kicks into gear, automatically extracting all the crucial details (Invoice Number, Date, Amount, etc.), saving the data to a Google Sheet, archiving the original file to Google Drive, and even replying with a friendly, AI-generated summary.

Say goodbye to manual data entry and lost receipts! 👋

✨ Key Features

🛠️ Architecture & Workflow

Here is a visual deep-dive into how the automation flows from start to finish:

Step-by-Step Process

1. Trigger: You send an image (invoice) to your Telegram Bot.
2. Download: The workflow downloads the image file from Telegram's servers to the n8n instance.
3. OCR Analysis: The image is sent to the OCR.space API, which returns the raw text found in the document.
4. Data Extraction: A custom code node parses the raw text to find key information (Invoice #, Date, Amount, etc.).
5. Database Update: The extracted data is added as a new row in your Google Sheet.
6. File Storage: The original image is uploaded to a specific folder in your Google Drive.
7. AI Processing: An AI Agent (LangChain) reviews the extracted info and constructs a helpful response.
8. Response: The bot replies to you on Telegram with a summary and a link to the database.

📋 Prerequisites

Before you begin, ensure you have the following:

• n8n Instance: You need n8n installed and running (Self-hosted or Cloud).
• Telegram Account: To create a bot.
• Google Account: For Sheets and Drive access.
• OCR.space API Key: Free key available here.
• OpenAI API Key: For the AI agent functionality.

⚙️ Installation & Setup

Follow these detailed steps to get your agent up and running.

1. Telegram Bot Setup

1. Open Telegram and search for @BotFather.
2. Send the command /newbot.
3. Follow the prompts to name your bot and get your HTTP API Token.
4. Save this token; you will need it for n8n credentials.

2. Google Cloud Setup

1. Google Cloud Console: Create a new project and enable Google Sheets API and Google Drive API.
2. Credentials: Create OAuth2 credentials (Client ID and Secret).
   • Note: For n8n, you'll need to set the redirect URI to your n8n instance's callback URL.
3. Google Sheet: Create a new Google Sheet.
   • Name the first sheet Sheet1.
   • Create the following headers in the first row:
     • Invoice Number
     • Date
     • Total Amount ($)
     • Billing Address
     • Due Date
     • Notes
4. Google Drive: Create a folder named Invoices (or whatever you prefer) to store the images.

3. OCR.space Setup

1. Visit OCR.space.
2. Register for a free API key.
3. You will receive the key via email.

4. OpenAI Setup

1. Go to platform.openai.com.
2. Generate a new API Key.
3. Ensure you have credits/billing set up.

5. n8n Configuration

1. Import Workflow:
   • Copy the JSON content of the workflow file.
   • In n8n, go to Workflows > Import from JSON and paste the code.
2. Configure Credentials:
   • Telegram Trigger: Create new credentials using your Bot Token.
   • Google Sheets: Connect your Google account using OAuth2.
   • Google Drive: Connect the same Google account.
   • OpenAI: Add your OpenAI API Key.
3. Update Node Settings:
   • Update Database Node: Select your specific Google Sheet and Sheet Name.
   • Add Invoice Image to Drive Node: Select your specific Google Drive folder.
   • Analyze Image Node: In the Header Parameters, replace YOUR API KEY HERE with your actual OCR.space API key.
4. Activate: Toggle the workflow to Active.

🚀 Usage Guide

Using the agent is incredibly simple:

1. Open your Telegram app.
2. Navigate to your new bot.
3. Upload a photo of an invoice or forward a file (PDF/Image) to the bot.
4. Wait a few seconds... ⏳
5. Ding! 🔔 You will receive a reply confirming the data extraction, along with a link to your dashboard.

Tip: Ensure the image is clear and the text is legible for the best OCR results.

💾 Data Structure & Customization

The system is designed to extract the following fields. You can modify the Parse Text node (Javascript) to add more fields if needed.

Adding New Fields

To add a new field (e.g., "Tax Amount"):

1. Open the Parse Text node in n8n.
2. Add a new line to the parsedData object:

   javascript
   taxAmount: text.match(/Tax:\s*([\d,.]+)/)?.[1],
   

3. Add a new column to your Google Sheet.
4. Update the Update Database node to map the new field.

🔧 Troubleshooting & FAQ

❌ The bot isn't responding!

• Check if the workflow is Active in n8n.
• Verify your Telegram credentials are correct.
• Check the n8n Executions tab for any error logs.

❌ OCR is missing data!

• The image might be blurry. Try a higher quality image.
• The regex patterns in the "Parse Text" node might need adjustment to match your specific invoice format.

❌ Google Drive/Sheets errors!

• Ensure your OAuth2 credentials haven't expired.
• Verify that the specific Sheet ID and Folder ID in the nodes match your actual files.

❓ Can I use a different OCR provider?

• Yes! You can replace the OCR.space node with Google Cloud Vision, AWS Textract, or any other OCR service supported by n8n.

🤝 Contributing

Contributions are welcome! If you have ideas for improvements, feel free to fork this repository and submit a pull request.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📄 License

Distributed under the MIT License. See LICENSE for more information.

Made with ❤️ by the Ai agent bros team

Website •
Documentation •
Support