> ## Documentation Index
> Fetch the complete documentation index at: https://docs.documind.cloud/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction to Documind

> Intelligent document extraction for automation workflows

## What is Documind?

Documind is an AI-powered document extraction platform that transforms unstructured documents into structured data. Whether you're processing invoices, forms, receipts, or contracts, Documind uses advanced LLM models to extract information accurately and efficiently.

## Key Features

<CardGroup cols={2}>
  <Card title="Multi-Model Extraction" icon="brain">
    Choose from Basic (single-model), VLM-based, or Advanced (multi-model ensemble) extraction modes for different accuracy and cost trade-offs.
  </Card>

  <Card title="Flexible Schemas" icon="diagram-project">
    Define custom schemas, use predefined templates, or let Documind generate schemas automatically from sample documents.
  </Card>

  <Card title="Confidence Scoring" icon="gauge-high">
    Every extracted field includes confidence scores. Low-confidence fields are automatically flagged for human review.
  </Card>

  <Card title="Review Workflow" icon="user-check">
    Built-in review system enables human-in-the-loop validation for critical data, ensuring accuracy when it matters most.
  </Card>

  <Card title="Credit-Based Pricing" icon="coins">
    Transparent per-page pricing with different costs for each extraction mode. No surprises.
  </Card>

  <Card title="Developer-First API" icon="code">
    RESTful API with comprehensive documentation, code examples, and SDKs for seamless integration.
  </Card>
</CardGroup>

## How Documind Works

<Steps>
  <Step title="Upload Documents">
    Upload PDF, Word, or image files to Documind. Each document receives a unique ID for tracking.
  </Step>

  <Step title="Define What to Extract">
    Specify the data you need using JSON Schema format. Use predefined schemas, generate from samples, or create custom ones.
  </Step>

  <Step title="Choose Extraction Mode">
    Select the extraction mode that fits your needs:

    * **Basic**: Fast single-model extraction
    * **VLM**: Vision-optimized for image-heavy documents
    * **Advanced**: Multi-model ensemble with confidence scoring
  </Step>

  <Step title="Get Structured Data">
    Receive extracted data as JSON with confidence scores. Low-confidence fields are flagged for review if using Advanced mode.
  </Step>

  <Step title="Review if Needed">
    For extractions flagged for review, humans verify and correct the data. Reviewed results replace initial extractions.
  </Step>
</Steps>

## Extraction Modes Comparison

| Feature               | Basic                | VLM            | Advanced             |
| --------------------- | -------------------- | -------------- | -------------------- |
| **Speed**             | Fastest              | Fast           | Moderate             |
| **Cost per page**     | 2-6 credits          | 10 credits     | 15 credits           |
| **Confidence scores** | No                   | No             | Yes                  |
| **Review flagging**   | No                   | No             | Yes                  |
| **Best for**          | Simple documents     | Scanned images | Complex forms        |
| **Models**            | Single (your choice) | Multiple VLMs  | Multi-model ensemble |

## Common Use Cases

### Invoice Processing

Extract line items, totals, vendor information, dates, and payment terms from invoices in any format.

### Form Digitization

Convert paper forms, applications, and surveys into structured database entries.

### Receipt Management

Pull amounts, merchants, dates, and categories from receipts for expense tracking.

### Contract Analysis

Extract key terms, parties, dates, and obligations from legal documents.

### Identity Verification

Extract information from IDs, passports, and verification documents.

## Architecture Overview

```
┌─────────────┐
│   Upload    │  POST /api/v1/upload
│  Documents  │  → Returns document_id(s)
└──────┬──────┘
       │
       ▼
┌─────────────┐
│   Extract   │  POST /api/v1/extract/{document_id}
│    Data     │  → Returns results + needs_review flag
└──────┬──────┘
       │
       ├──► needs_review = false
       │    ✓ Use results immediately
       │
       └──► needs_review = true
            ↓ Human reviews and corrects
            ↓ PUT /api/v1/review/{document_id}
            ↓
            ✓ Poll GET /api/v1/data/extractions?document_id=...
              until is_reviewed = true
```

## Getting Started

<CardGroup cols={2}>
  <Card title="Quick Start" icon="rocket" href="/guides/quickstart">
    Get started with your first extraction in 5 minutes
  </Card>

  <Card title="Schema Design" icon="pencil" href="/guides/schema-design">
    Learn how to create effective extraction schemas
  </Card>

  <Card title="API Reference" icon="book" href="/api-reference/introduction">
    Explore all available endpoints and parameters
  </Card>

  <Card title="Tutorials" icon="graduation-cap" href="/guides/tutorials/invoice-processing">
    Step-by-step tutorials for common use cases
  </Card>
</CardGroup>

## Next Steps

1. **[Try the Quick Start Guide](/guides/quickstart)** - Extract your first document in 5 minutes
2. **[Explore Use Case Tutorials](/guides/tutorials/invoice-processing)** - Learn from real-world examples
3. **[Design Effective Schemas](/guides/schema-design)** - Master schema creation for better results
4. **[Integrate with Your App](/api/quickstart)** - Build production-ready integrations