Documentation link:
https://docs.nocodeventure.com/ocr-with-claude-ai
Claude Vision Plugin Guide
📘 This plugin extracts text and structured data from images or PDFs using Claude AI and returns structured JSON. You can optionally convert PDFs to images before sending.
🔐 API Key Setup
You must provide a Claude API key.
Option 1: Set the key in the plugin settings under "Claude API Key"
Option 2: Use the api_key field in the workflow action to override the default key
Get your key at:
https://console.anthropic.com/settings/keys📥 Supported Input
Images: PNG JPG JPEG WebP
PDFs: Any valid PDF file
File URLs must be full https URLs
🧠 Prompt Options
system_prompt: Define Claude’s role (example: You are a document parser)
user_prompt: Ask what to extract (example: Extract all fields as JSON)
extract_json: Yes or No – if Yes, Claude response is parsed into structured JSON
📤 Outputs
text: Full Claude response
model: Claude model used
id: Claude message ID
extracted_json: JSON string if parsing is enabled
⚙️ Server-side Actions
1️⃣ Send File to Claude
Inputs: file_url, api_key, system_prompt, user_prompt, extract_json
Returns: text, model, id, extracted_json
2️⃣ Send Images to Claude
Inputs: list of image URLs or base64 strings, system_prompt, user_prompt, extract_json
Returns: text, model, id, extracted_json
📄 Element – PDF to Image Converter
Inputs:
pdf: PDF file URL
start_page: First page to convert
end_page: Last page to convert
scale: Render scale (default is 1)
upload_to_bubble: Yes to upload images, No to return base64 only
States:
images: List of image URLs or base64 strings
page_count: Total pages in PDF
Events:
saved: Triggered when all images are ready
Actions:
Convert all PDF pages to images
Convert page range to images
🧾 Use this plugin to extract structured data from scanned forms, invoices, receipts, and more using Claude AI.
Demo page:
https://ncv-demos4.bubbleapps.io/version-test/claude-ocrEditor:
https://bubble.io/page?id=ncv-demos4&tab=Design&name=claude-ocr&type=page&wf_item=cmQnV