Airtable OCR

This guide shows you how to extract text from images and scanned documents in Airtable using AI-powered OCR. You'll learn to process receipts, business cards, handwritten notes, screenshots, and any other image-based text using Anthropic's Claude AI through the Data Fetcher extension.

The Airtable OCR integration works with photos, scanned PDFs, screenshots, and even handwritten text. There's no coding required, and you can set it up to run automatically whenever new images are added to your base.

Data Fetcher's Anthropic integration makes this process simple and code-free, using Claude's advanced vision capabilities to read and understand text in your images.

Set Up Your Documents Table

To get started, we'll create a table for your files and configure a view to track which items need OCR processing.

1. Create a new table called "Documents".

2. Add an Attachment field called "File" (or use your existing attachment field).

3. Upload your files to the "File" field. You can use receipts, business cards, PDFs, screenshots, photos of whiteboards, or any images and documents containing text.

4. Create a new field for the extracted data:

• Extracted Text (Long text field) - for the full text content.

5. Create a new Grid view called Needs OCR with the following filters:

• File - is not empty.
• Extracted Text - is empty.

This filtered view ensures we only process files that haven't been analyzed yet, saving API credits and preventing duplicate processing.

Generate an Anthropic (Claude) API Key

Before connecting to Claude's vision API in Airtable, you'll need to create an API key.

1. Sign up for an Anthropic Console account or log in if you already have one.

2. Navigate to API Keys in your account settings.

3. Click Create Key.

4. Select a workspace and give your key a name like "Data Fetcher" and click Add.

5. Copy the API key and save it somewhere secure. You won't be able to see it again.

Note: You'll need to set up billing and purchase some credits in your Anthropic account to use the API.

Install the Airtable OCR Integration

1. Add the Data Fetcher extension to your Airtable base.

2. Sign in to Data Fetcher or sign up if you don't have an account.

3. Once you're logged in, click Create your first request.

Configure OCR with Anthropic (Claude) API

We'll use Anthropic's Claude AI for OCR because it handles both images and documents through the same workflow. Unlike OpenAI (which requires different endpoints for images vs PDFs) or Airtable's built-in "Analyze attachments" field agent (which doesn't support images at all), Claude processes everything in one request.

Follow these steps to set up your Airtable OCR integration:

1. Select Anthropic under Application.

2. Rename the request to "Extract text".

3. Copy and paste your Anthropic API key under Authorization.

4. Under Endpoint, select Create a message.

5. Click Save and Continue in the bottom right corner to proceed to the next section.

6. Select Claude 4 Sonnet under Model.

Next, we'll configure Claude to extract text from your images and documents.

Send Messages to the Claude API

1. Copy and paste the following text under Messages:

Extract all text from this image or document. Return only the extracted text, no extra commentary.

Now we'll reference the files from your table.

2. Click the + button next to Messages.

3. In the dialog that opens:

• Select File under Field
• Under Run for every record in view, select Needs OCR

4. Click Confirm to save and close the dialog.

You should now see the "File" field referenced in the message.

Extract Structured Data (Optional)

If your documents contain consistent fields (like invoices or business cards), you can extract specific data points. Replace the text in the message above with:

Extract data from this document and return as JSON:
{
  "name": "",
  "email": "",
  "phone": "",
  "company": "",
  "documentNumber": ""
}
If a field is not found, leave it empty. Return only the JSON, no extra text.

This approach works well for business cards, receipts, invoices, or any documents with standard fields.

5. Click Save and Run to proceed to the next section.

Map the Extracted Text to Your Table

After the Airtable OCR integration runs, the Response Field Mapping screen will open. Here, you'll configure how the extracted text is imported into your table.

The response contains one field, "Message", which holds all the text Claude extracted from your files.

1. Click Existing field under Message, then select Extracted Text.

2. Click Save and Run.

Return to your default grid view to see the extracted text from all your images and documents.

Mapping Structured Data

If you used the JSON extraction method, you'll see multiple fields in the response mapping:

Map each JSON field to corresponding fields in your table:

• Json name → Name field
• Json email → Email field
• Json phone → Phone field

This lets you automatically organize extracted data into structured Airtable fields.

Run Airtable OCR Automatically

To automatically extract text whenever new files are added to your table, you can use Data Fetcher's Triggers feature.

Triggers are a premium feature, so follow these steps to upgrade and set up automation:

1. Open the request in Data Fetcher and scroll down to the Schedule / Trigger / Webhook URL tabs.

2. Select Trigger and click Upgrade.

3. Choose a paid plan that fits your needs and complete the payment process.

4. Click + Authorize to give Data Fetcher access to your Airtable base.

5. Click I understand, let's Authorize in the warning dialog.

6. In the authorization window, click Add a base, then select All all resources.

7. Click Grant Access.

Now you can create the trigger:

1. Select Record created.

2. Select Documents under Table and Needs OCR under View.

3. Click Save at the bottom of the screen.

Your Airtable OCR automation is now active! Whenever you upload new images or documents, Data Fetcher will automatically extract the text without any manual work.

Conclusion: Automated OCR in Airtable

You've successfully set up an automated OCR system in Airtable using Claude AI. Your workflow now automatically extracts text from any image or document you upload, saving hours of manual data entry.

Common Use Cases for Airtable OCR

This OCR workflow is particularly useful for:

• Receipt Management: Extract vendor names, dates, amounts, and line items from receipts and invoices for expense tracking and accounting automation.
• Business Card Scanning: Convert stacks of business cards into structured contact records with names, companies, emails, and phone numbers.
• Document Digitization: Transform paper archives, contracts, and legacy documents into searchable digital text in your Airtable base.
• Form Processing: Extract data from completed forms, applications, or surveys that were filled out by hand or printed.
• Meeting Notes: Convert whiteboard photos or handwritten meeting notes into actionable text that can be searched and organized.
• Screenshot Documentation: Capture text from software screenshots, error messages, or social media posts for analysis and record-keeping.

Frequently Asked Questions

What file types does this OCR integration support?
The Airtable OCR integration works with common image formats (JPEG, PNG, GIF, WebP) and PDF documents. Claude can read both digital and scanned PDFs, as well as photos of physical documents.

How accurate is the text extraction?
Claude's vision capabilities provide high accuracy even with challenging inputs like handwritten text, curved receipts, or low-quality scans. For best results, ensure images are well-lit and text is clearly visible.

Can I extract data in different languages?
Yes, Claude supports OCR in multiple languages including English, Spanish, French, German, Japanese, Chinese, and many others. Simply upload documents in any supported language.