Extractify
Sign InGet Started

Documentation

Everything you need to master Extractify.

Getting Started

What is Extractify?

Extractify automatically extracts structured data from your documents (invoices, forms, receipts, etc.) using enterprise-grade OCR and document parsing engines. Upload a document, and we'll extract key-value pairs like names, dates, amounts, and more.

✨
TIP
New to document processing? Start with a simple invoice or receipt for best results!

Your First Document

1. Sign in to your account
2. Navigate to Dashboard
3. Click Upload Document
4. Select a PDF file (max size depends on your plan)
5. Click Process
6. Wait for processing to complete
7. View your results in the Results page
Document upload interface showing drag-and-drop area
Document upload interface showing drag-and-drop area
💡
NOTE
Processing typically takes 10-30 seconds depending on document size and complexity.

Processing Documents

Supported File Types

  • PDF documents (single or multi-page)
  • Maximum pages per file: varies by plan (Free: 25, Starter: 50, Pro: 100+)

    • Processing Limits

    Your plan determines:

  • Pages per file: How many pages can be in one document
  • Monthly pages: Total pages you can process per month
  • Storage: How many results you can keep

    • Check your current usage in Settings → Billing.

    Batch Processing

    Upload multiple files at once:

    1. Click Upload Document
    2. Select multiple files (Ctrl+Click or Cmd+Click)
    3. All files will be queued for processing
    4. View progress in the Results page

    Viewing Results

    View Modes

    #### 📊 List View (Default)

    Shows extracted data as a table with one row per field:

  • Field: The label (e.g., "Invoice Number")
  • Value: The extracted data (e.g., "INV-12345")
  • Confidence: AI confidence score (0-100%)

    • Best for: Reviewing all extracted data, editing values

    Results page showing extracted data in table format
    Results page showing extracted data in table format
    ✨
    TIP
    Green badges (95%+) indicate high confidence. Yellow badges (80-94%) may need verification.

    #### 🔄 Transformed View

    Groups data by page and pivots fields into columns:

  • Each row represents one page
  • Columns are the field names
  • Values are filled in for each page

    • Best for: Multi-page documents, exporting to Excel

    #### 📄 Text View

    Shows the raw OCR text extracted from the document:

  • Full unstructured text
  • Useful for finding data the Form Parser missed
  • Required for Click-to-Fill feature

    • Best for: Manual verification, finding missing data

    #### 💻 Raw JSON View

    Shows the complete API response in JSON format:

  • For developers and advanced users
  • Includes all metadata and confidence scores

    • Data Cleaning Tools

    Overview

    Data cleaning tools automatically normalize your extracted data for better export quality.

    🔴
    IMPORTANT
    Data cleaning is available on Starter and Pro plans only.

    Available Tools

    #### 📅 Clean Dates

    Converts all date formats to ISO standard (YYYY-MM-DD):

  • "Oct 12, 2023" → "2023-10-12"
  • "10/12/2023" → "2023-10-12"
  • "10/12/23" → "2023-10-12"

    • How to use:

    1. View results in List mode
    2. Click 📅 Dates button
    3. All dates are instantly normalized

    #### 💰 Clean Currency

    Removes currency symbols and formats numbers:

  • "$1,234.56" → "1234.56"
  • "($500.00)" → "-500.00"
  • "€1.234,56" → "1234.56"

    • How to use:

    1. View results in List mode
    2. Click 💰 Currency button
    3. All currency values are cleaned
    Before and after comparison of data cleaning
    Before and after comparison of data cleaning

    #### ✨ Clean All

    Applies both date and currency cleaning in one click.

    #### ↩️ Undo

    Reverts your last cleaning operation.

    💡
    NOTE
    Cleaning is temporary - refresh the page to restore original data.

    Click-to-Fill Feature

    What is Click-to-Fill?

    Quickly fill empty or incorrect cells by selecting text directly from the OCR output. Perfect for fixing missing data or correcting errors.

    ✨
    TIP
    Click-to-Fill is available on Starter and Pro plans.

    How to Use

    Step 1: Activate a Cell

    1. Go to Results page
    2. Switch to List view
    3. Click any cell you want to fill
    4. The cell gets a blue border
    5. You're automatically switched to Text view

    Step 2: Build Your Value

    1. You'll see a blue banner: "Click words to build value for: [Field Name]"
    2. Click words in the OCR text
    3. Each word is added to the Preview box
    4. Example: Click "John" then "Smith" → Preview shows "John Smith"
    Click-to-Fill interface showing cell selection and word picking
    Click-to-Fill interface showing cell selection and word picking

    Step 3: Confirm

  • ✓ Confirm (green button): Fills the cell and returns to List view
  • 🗑️ Clear (gray button): Clears your selection, start over
  • Cancel (white button): Exits without filling

    • Tips

  • Click words in the order you want them
  • Use the preview to verify before confirming
  • Press ESC to cancel anytime
  • Perfect for split values like names or addresses

    • Example Use Cases

  • Missing name: Click "John" + "Smith" → Confirm
  • Split address: Click "123" + "Main" + "Street" → Confirm
  • Wrong value: Click correct text from OCR → Confirm

    • Exporting Data

    Export Formats

    #### CSV Export

  • Available on all plans
  • Opens in Excel, Google Sheets
  • Row limit: varies by plan (Free: 10 rows, Starter/Pro: unlimited)

    • #### TXT Export

  • Available on all plans
  • Downloads the raw OCR text extracted from the document
  • Best for manual review and copy-pasting

    • How to export:

    1. Go to the Documents page or Results dashboard.
    2. For a single file: Click the TXT or CSV button on the file card or row.
    3. For all files: Click Download All in the Results dashboard (Pro feature).

    #### JSON Export

  • Available on Starter and Pro plans
  • For developers and advanced users
  • Includes all metadata

    • #### Excel Export

  • Coming soon (Pro plan)

    • #### Power BI / Tableau

  • Coming soon (Pro plan)

    • Export Limits

  • Free: 10 rows per export
  • Starter: Unlimited rows
  • Pro: Unlimited rows + advanced formats

    • Managing Your Account

    Viewing Usage

    1. Go to Settings → Billing
    2. See your current usage:

    - Pages processed this month

    - Remaining pages

    - Storage used

    Upgrading Your Plan

    1. Go to Settings → Billing
    2. Click Upgrade Plan
    3. Choose Starter or Pro
    4. Enter payment details
    5. Instant activation

    Plan Comparison

    Feature
    Free
    Starter
    Pro
    Monthly Pages
    100
    2,500
    10,000
    Max pages/file
    25
    50
    100+
    Results storage
    25
    500
    1,000
    Export rows
    10
    Unlimited
    Unlimited
    Data cleaning
    ❌
    ✅
    ✅
    Click-to-Fill
    ❌
    ✅
    ✅
    Search/Filter
    ❌
    ✅
    ✅

    Troubleshooting

    Common Issues

    "Processing failed"

  • Check file size (must be under your plan's limit)
  • Ensure file is a valid PDF
  • Try re-uploading

    • "No data extracted"

  • Document may be an image-only PDF (OCR will still extract text)
  • Try a clearer scan or higher quality PDF
  • Check the Text view to see if OCR worked

    • "Low confidence scores"

  • Poor scan quality
  • Handwritten text
  • Unusual document format
  • Use Click-to-Fill to manually correct

    • "Export limited to 10 rows"

  • You're on the Free plan
  • Upgrade to Starter or Pro for unlimited exports

    • "Can't use Data Cleaning"

  • Feature requires Starter or Pro plan
  • Upgrade in Settings → Billing

    • Getting Help

  • Email: support@yourdomain.com
  • In-app chat: Click the help icon
  • Documentation: docs.yourdomain.com

    • Keyboard Shortcuts

    Shortcut
    Action
    ESC
    Cancel Click-to-Fill mode
    Ctrl/Cmd + S
    Download CSV (when viewing results)

    Best Practices

    For Best Results

    1. Use high-quality scans (300 DPI or higher)
    2. Ensure text is readable (not blurry or skewed)
    3. Process similar documents together (invoices, receipts, etc.)
    4. Review confidence scores (below 80% may need verification)
    5. Use Click-to-Fill for missing or incorrect data

    Workflow Recommendations

    1. Upload documents
    2. Review results in List view
    3. Use Click-to-Fill to fix any errors
    4. Apply Data Cleaning (dates, currency)
    5. Switch to Transformed view for multi-page docs
    6. Export to CSV/Excel

    How Document Processing Works

    Processing Overview

    Files are securely uploaded and processed using a document extraction engine optimized for forms and structured data. Results are returned as downloadable CSV and JSON files.

    Processing Steps:

    1. Document upload (encrypted in transit)
    2. OCR and form field detection
    3. Data extraction and validation
    4. Confidence scoring for each field
    5. Results delivered in multiple formats
    💡
    NOTE
    Processing typically takes 10-30 seconds depending on document complexity and page count.

    Technology & Models

    Extractify uses a production-grade document extraction engine designed for forms and structured documents. We may update or improve the underlying engine over time to improve accuracy, speed, and cost-efficiency.

    Current Implementation:

  • We currently use Google Document AI under the hood for document parsing
  • Combined with proprietary data cleaning and transformation tools
  • Optimized for invoices, receipts, forms, and structured documents
    • ✨
    TIP
    Specific engine details and API specifications are available upon request for enterprise customers.

    AI Data Refinement

    What is AI Refinement?

    Use advanced AI to clean, format, and enrich your data beyond standard cleaning tools. You can write natural language instructions (e.g., "Normalize company names" or "Translate description to English") to transform your data.

    Using the AI Workbench

    1. Open a file in the Results page.
    2. Click the AI Lab tab (or Refine button).
    3. Type your instruction in the text box.
    4. Click Run AI Refinement.

    Targeted Processing

    To save credits and be more precise, you can refine only specific rows:

    1. Toggle Process Selected Only to ON.
    2. In the Live Data Preview table, check the boxes next to the rows you want to change.
    3. The AI will *only* modify the selected rows; others remain untouched.

    History & Safety

  • Undo: Made a mistake? Click Undo in the Action History panel to revert the last step.
  • Retry: If an action fails, use the Retry button to try again.
  • Reset: Use Reset to Original to discard all unsaved changes and start over.

    • Saving Your Work

    Refinements are temporary until saved!

  • Click Save Version to create a deeper copy of your refined dataset in the cloud.
  • You can give it a custom name (e.g., "Cleaned Q4 Reports").

    • Billing & Subscriptions

    Managing Your Plan

    #### Upgrading or Downgrading

    You can change your plan at any time:

    1. Go to Settings → Billing
    2. Click Upgrade on a higher plan or Downgrade on a lower plan
    3. Changes take effect immediately (prorated charges may apply)

    #### Cancelling Your Subscription

    1. Go to Settings → Billings (or just Settings)
    2. Click Cancel Subscription
    3. Confirm your choice
    💡
    NOTE
    Grace Period: When you cancel, you retain access to all plan features until the end of your current billing period. Your plan status will change to "Cancellation Scheduled".

    Regaining Access

    If you change your mind during the grace period (before your plan expires):

    1. Go to Settings
    2. Click the Regain Access link in the "Cancellation Scheduled" banner
    3. You will be redirected to the secure billing portal where you can click "Don't cancel subscription"
    4. Your plan will resume normal renewal—no data lost!

    Plans, Limits & Usage

    Per-Plan Limits

    Plan
    Max Pages/File
    Monthly Pages
    Storage
    PAYG Option
    Free
    25
    100
    25 results
    No
    Starter
    50
    1,000
    100 results
    Yes
    Pro
    100+
    10,000+
    Unlimited
    Yes

    What Happens When Limits Are Hit?

    File Upload Limit:

  • If a file exceeds your max pages per file, you'll see an error before processing
  • Solution: Upgrade your plan or split the document

    • Monthly Page Limit:

  • Processing is blocked when you hit your monthly limit
  • Solution: Wait for next billing cycle or purchase additional credits

    • Storage Limit:

  • You cannot process new documents when storage is full
  • Solution: Delete old results or upgrade your plan

    • Privacy & Data Handling

    Data Security

  • Documents are encrypted in transit (TLS) and at rest
  • Processed on secure cloud infrastructure
  • Access controls and authentication required

    • Data Usage

    Uploaded documents are processed for extraction purposes only and are not used to train public models.

  • We never share your data with third parties
  • You can delete results anytime from the Results page
  • Data retention varies by plan (Free: 7 days, Starter: 30 days, Pro: unlimited)

    • Your Control

  • Download your data anytime (CSV, JSON, Excel)
  • Delete individual results or bulk delete
  • Export before deletion for permanent backup

    • Need more help? Contact support@extractifyhq.com