Description

n8n-nodes-ocr-ai

This is an n8n community node for document extraction using AI. It wraps the ocr-ai library (v1.0.3) to provide OCR and structured data extraction capabilities.

Features

• Multi-provider support: Gemini, OpenAI, Claude, Grok, and Vertex AI
• Text extraction: Extract plain text from documents
• JSON extraction: Extract structured data matching a custom schema
• Multiple input types: Binary data, URL, or Base64
• Supported formats: PDF, images (jpg, png, gif, webp, bmp, tiff), text files

Installation

Community Nodes (Recommended)

1. Go to Settings > Community Nodes
2. Select Install
3. Enter n8n-nodes-ocr-ai in the package name field
4. Accept the risks and click Install

Manual Installation

cd ~/.n8n/nodes
pnpm install n8n-nodes-ocr-ai

Credentials

Configure credentials based on the provider you want to use:

| Provider | Credential Type | Required Fields |
|———-|—————-|—————–|
| Gemini | OCR AI Gemini API | API Key |
| OpenAI | OCR AI OpenAI API | API Key |
| Claude | OCR AI Claude API | API Key |
| Grok | OCR AI Grok API | API Key |
| Vertex AI | OCR AI Vertex AI | Project ID, Location |

Usage

Extract Text

1. Add the OCR AI node to your workflow
2. Select a provider and configure credentials
3. Choose Extract Text operation
4. Select input type (Binary, URL, or Base64)
5. Configure the input source
6. Run the workflow

Extract JSON

1. Add the OCR AI node
2. Choose Extract JSON operation
3. Define a JSON schema describing the structure you want to extract
4. The node will return structured data matching your schema

#### Example Schema

{
  "invoice_number": "string",
  "date": "string",
  "total": "number",
  "items": [{
    "description": "string",
    "quantity": "number",
    "price": "number"
  }]
}

Options

| Option | Description |
|——–|————-|
| Custom Prompt | Guide the extraction with a custom prompt |
| Language | Set extraction language (default: auto) |
| Model | Override the default model for the provider |
| Temperature | Control randomness (0-2) |
| Max Tokens | Maximum tokens in response |

Default Models

| Provider | Default Model |
|———-|————–|
| Gemini | gemini-1.5-flash |
| OpenAI | gpt-4o |
| Claude | claude-sonnet-4-20250514 |
| Grok | grok-2-vision-1212 |
| Vertex AI | gemini-2.0-flash |

Development

Install dependencies
pnpm install
Build
pnpm build
Development mode
pnpm dev
Lint
pnpm lint

License

MIT

Dependencies

• ocr-ai v1.0.3