Process documents with SFTP - ABBYY Documentation

Overview

ABBYY Vantage enables automated document processing with intelligent extraction capabilities. This guide focuses on the SFTP-based workflow, which allows you to upload documents to a shared folder and automatically retrieve processed results. What you’ll accomplish:

Create and configure a Process Skill
Set up SFTP access for input and output
Upload documents via command line
Download extraction results

Time to complete: 15-20 minutes

Step 1: Create a Process Skill

A Process Skill is the foundation of your document processing workflow in Vantage.

Navigate to the Documents section in ABBYY Vantage.
Click on Skill Designer in the left sidebar.
Select Process Skill from the skill types.

In the Create Process skill dialog, enter:
- Skill Name: A descriptive name for your process (e.g., “Process Skill Test 1”)
- Description: Optional description of what this skill does
- Technology Core Version: 3.0 (recommended)

Click Create

Step 2: Configure the Process Skill Workflow

After creating your skill, you’ll set up the document processing flow.

In the Skill Designer, you’ll see the Select Process Skill panel on the right
Choose Simple Document skill workflow for processing single document types

This creates a pre-configured workflow with:

Input connector: Where documents enter the system
OCR skill: Optical character recognition
Extract skill: Data extraction
Output connector: Where results are delivered

The workflow will appear in the center canvas showing: Input → OCR → Extract → Output

Step 3: Set Up Input via SFTP

Configure the Input activity to accept documents via SFTP.

Click on the Input activity in the workflow.
In the Actions panel on the right, check Select additional source.
Select the Shared folder radio button.
Click Settings to configure.

Configure Input Settings

In the Input Settings: Shared Folder dialog:

Folder settings:

Folder name: Input (default)
Import method: Single file

Credentials for import (displayed for your use):

Folder link: .../Input
Exceptions folder link: .../Input-exception
Username: A generated UUID (e.g., 685df03c-9e22-41cb-a96f-4c408ab52735)
Password: Displayed with show/hide option

Copy these credentials — you’ll need them to connect via SFTP. Files that can’t be imported will be moved to the exceptions folder. Vantage checks for files every 20 seconds after publishing the Skill.

Click Save.

Step 4: Set Up Output via SFTP

Configure the Output activity to deliver results via SFTP.

Click on the Output activity in the workflow.
In the Actions panel, check Select additional destination.
Select the Shared folder radio button.
Click Settings to configure.

Configure Output Settings

In the Output Settings: Shared Folder dialog:

Info:

Folder name: Output (default)
Folder: .../Output
Username: A generated UUID (e.g., 685df03c-9e22-41cb-a96f-4c408ab52735)
Password: Displayed with show/hide option

Exported Data:

Fields (JSON): Default export format

Results are organized in subfolders named after the source file and transaction ID (e.g., invoice_1.jpg-4ac2a31b-d1b6-4231-99b2-39f00bc5df82). Files are automatically removed according to the data retention policy.

Click Save.

Step 5: Configure Extraction Skills

Set up which document types the Extract activity should recognize and process.

Click on the Extract activity in the workflow.
In the Actions panel, you’ll see Document Skills.

Available document skills include:

Invoice
Invoice AU-NZ
Invoice CA
Invoice ES
Invoice UA
Invoice US

Select the document types relevant to your use case.
Click Add Skill if you need additional document types.
Use Edit Mapping to customize field extraction if needed.

Step 6: Publish and Get Your Skill ID

Before you can use SFTP, you need to publish the skill and locate its ID.

Click the Publish button in the top right of the Skill Designer.
Once published, note the Skill ID from the browser URL.

The URL format is:

https://vantage-preview.abbyy.com/skill-editor/Processing/{skill-id}

For example: fc7116ea-ab6c-40fa-b10d-1aef61aaaa2c

Keep your Skill ID handy — you’ll need it to access the correct Input and Output folders in SFTP.

Step 7: Connect to SFTP

Now you’re ready to connect via SFTP using the credentials from your Input/Output settings.

Connection Details

From your Input/Output settings, you have:

Server: 172.172.228.138
Port: 2022
Username: Your generated UUID
Password: Your generated password
Base Path: /{skill-id}/

Connect via Terminal

Open your terminal and connect using:

sftp -P 2022 {username}@172.172.228.138

Example:

sftp -P 2022 8d1bf446-3abb-4ddb-a8a8-ed87fac7f171@172.172.228.138

Use capital -P for the port number in SFTP (lowercase -p is used for a different purpose).

When prompted, enter your password.

Step 8: Navigate the SFTP Folder Structure

Once connected, you’ll see the SFTP prompt: sftp>

View Available Skills

List all available Process Skills:

sftp> ls

You’ll see folders for each skill (identified by their skill IDs), plus:

Catalogs - Document skill catalogs
Catalogs-exception - Failed catalog imports
skills-import - Skill import directory

Navigate to Your Skill

Use the skill ID from Step 6:

sftp> cd {your-skill-id}
sftp> ls

You should see three folders:

Input - Where you upload documents for processing
Input-exception - Failed imports are moved here
Output - Where processed results appear

Step 9: Upload Documents for Processing

Navigate to the Input folder and upload your documents.

Navigate to Input Folder

sftp> cd Input
sftp> pwd
Remote working directory: /{skill-id}/Input

Upload a Single File

Upload a file with the full path:

sftp> put ~/Documents/Sample\ Files/Demo\ Docs/invoice_1.jpg

Or change your local directory first:

sftp> lcd ~/Documents/Sample\ Files/Demo\ Docs
sftp> lls
sftp> put invoice_1.jpg

Upload Multiple Files

sftp> mput *.jpg
sftp> mput invoice_*.pdf

Files uploaded to Input are automatically deleted after successful processing (usually within 20 seconds). If you don’t see your file after uploading, it means Vantage has already picked it up for processing.

Verify Upload

Check if files are still in the Input folder:

sftp> ls

If the folder is empty or your file is missing, processing has begun.

Step 10: Check Processing Status

Monitor Input-exception Folder

If a file fails to import, it will appear here:

sftp> cd ..
sftp> cd Input-exception
sftp> ls

An empty folder means all files imported successfully.

Wait for Processing

Processing time varies based on:

Document complexity
Extraction skills configured
Current system load

Typically, simple invoices process within 30-60 seconds.

Step 11: Download Processed Results

Once processing completes, results appear in the Output folder.

Navigate to Output Folder

sftp> cd ..
sftp> cd Output
sftp> ls

Understanding Output Structure

Results are organized in subfolders with this naming pattern:

{original-filename}-{transaction-id}

Example:

invoice_1.jpg-4ac2a31b-d1b6-4231-99b2-39f00bc5df82

View Contents of a Result Folder

sftp> cd invoice_1.jpg-4ac2a31b-d1b6-4231-99b2-39f00bc5df82
sftp> ls

Typical contents:

Invoice_US.json - Extracted data in JSON format
Invoice_US.csv - Extracted data in CSV format (if configured)
Additional files depending on your output settings

Download Results

Set your local download directory:

sftp> lcd ~/Downloads
sftp> lpwd
Local working directory: /Users/yourname/Downloads

Download a single file:

sftp> get Invoice_US.json

Download all files in the folder:

sftp> mget *

Download the entire result folder recursively:

sftp> cd ..
sftp> get -r invoice_1.jpg-4ac2a31b-d1b6-4231-99b2-39f00bc5df82

Understanding the JSON Output

The default output is a JSON file containing:

Document structure - Layout and organization
Extracted field values - Data pulled from the document
Confidence scores - Reliability of each extraction
Rule errors - Any validation issues
Field metadata - Additional information about each field

Example structure:

{
  "Document": {
    "DocumentType": "Invoice US",
    "Fields": [
      {
        "Name": "InvoiceNumber",
        "Value": "INV-12345",
        "Confidence": 0.98
      },
      {
        "Name": "Total",
        "Value": "1250.00",
        "Confidence": 0.95
      }
    ]
  }
}

For detailed JSON schema documentation, refer to the ABBYY Vantage API documentation.

Useful SFTP Commands

Command	Description	Example
`ls`	List remote files	`ls`
`lls`	List local files	`lls`
`pwd`	Show remote directory	`pwd`
`lpwd`	Show local directory	`lpwd`
`cd folder`	Change remote directory	`cd Output`
`lcd folder`	Change local directory	`lcd ~/Downloads`
`put file`	Upload single file	`put invoice.pdf`
`mput files`	Upload multiple files	`mput *.jpg`
`get file`	Download single file	`get result.json`
`mget files`	Download multiple files	`mget *`
`get -r folder`	Download folder recursively	`get -r results/`
`exit`	Disconnect	`exit`

Troubleshooting

Connection Issues

Problem: “Connection refused” or timeout errors Solutions:

Verify you’re connected to the company VPN
Check the port number (should be 2022 with capital -P)
Confirm your username and password are correct
Test basic network connectivity: ping 172.172.228.138

Upload Issues

Problem: Files not being processed Solutions:

Check the Input-exception folder for failed imports
Verify file format is supported (PDF, JPG, PNG, TIFF, etc.)
Ensure the Process Skill is published
Check file size limits (varies by configuration)

No Output Files

Problem: Output folder is empty after uploading Solutions:

Wait longer - processing can take 30-60 seconds or more
Verify Output is configured with “Shared folder” destination
Check transaction logs in the Vantage web interface
Confirm the Extract activity has appropriate document skills configured

Permission Errors

Problem: “Permission denied” when uploading/downloading Solutions:

Verify you’re using the correct credentials for Input vs Output
Input credentials allow: upload, delete
Output credentials allow: view, download, rename, delete
Check that the skill is published and active

Best Practices

Security

Never share SFTP credentials in plain text
Use secure channels to transmit passwords
Rotate credentials periodically
Limit SFTP access to necessary personnel

File Management

Use descriptive filenames that include dates or IDs
Process files in batches during off-peak hours if possible
Download and archive results promptly
Clean up old result folders regularly

Monitoring

Check Input-exception folder regularly for failed imports
Monitor processing times to identify bottlenecks
Review extraction confidence scores in results
Set up alerts for critical failures (if available)

Workflow Optimization

Group similar document types together
Configure appropriate document skills for your use case
Test with sample documents before production use
Document your SFTP workflow for team reference

Next Steps

Now that you have a working SFTP-based document processing pipeline:

Test with various documents - Upload different file types and document variations
Review extraction accuracy - Check the JSON results and confidence scores
Customize extraction - Use Edit Mapping to fine-tune field extraction
Automate uploads - Create scripts to automate file uploads
Integrate with systems - Connect your applications via the Vantage REST API
Scale your workflow - Add more document skills or create additional Process Skills

For advanced configurations, API integration, or custom workflows, refer to the ABBYY Vantage documentation.

Additional Resources

REST API Documentation: For programmatic file uploads and result retrieval
Skill Catalog: Browse pre-built document skills
Support: Contact ABBYY support for technical assistance

Summary

You’ve successfully:

✅ Created a Process Skill in ABBYY Vantage
✅ Configured SFTP access for input and output
✅ Set up document extraction skills
✅ Connected via SFTP command line
✅ Uploaded documents for processing
✅ Downloaded extraction results

Your document processing pipeline is now operational and ready for production use!

Overview

Working with Skills

Processing Documents

Working with LLMs

Developer Tools

Documentation Index

​Overview

​Step 1: Create a Process Skill

​Step 2: Configure the Process Skill Workflow

​Step 3: Set Up Input via SFTP

​Configure Input Settings

​Step 4: Set Up Output via SFTP

​Configure Output Settings

​Step 5: Configure Extraction Skills

​Step 6: Publish and Get Your Skill ID

​Step 7: Connect to SFTP

​Connection Details

​Connect via Terminal

​Step 8: Navigate the SFTP Folder Structure

​View Available Skills

​Navigate to Your Skill

​Step 9: Upload Documents for Processing

​Navigate to Input Folder

​Upload a Single File

​Upload Multiple Files

​Verify Upload

​Step 10: Check Processing Status

​Monitor Input-exception Folder

​Wait for Processing

​Step 11: Download Processed Results

​Navigate to Output Folder

​Understanding Output Structure

​View Contents of a Result Folder

​Download Results

​Understanding the JSON Output

​Useful SFTP Commands

​Troubleshooting

​Connection Issues

​Upload Issues

​No Output Files

​Permission Errors

​Best Practices

​Security

​File Management

​Monitoring

​Workflow Optimization

​Next Steps

​Additional Resources

​Summary

Overview

Step 1: Create a Process Skill

Step 2: Configure the Process Skill Workflow

Step 3: Set Up Input via SFTP

Configure Input Settings

Step 4: Set Up Output via SFTP

Configure Output Settings

Step 5: Configure Extraction Skills

Step 6: Publish and Get Your Skill ID

Step 7: Connect to SFTP

Connection Details

Connect via Terminal

Step 8: Navigate the SFTP Folder Structure

View Available Skills

Navigate to Your Skill

Step 9: Upload Documents for Processing

Navigate to Input Folder

Upload a Single File

Upload Multiple Files

Verify Upload

Step 10: Check Processing Status

Monitor Input-exception Folder

Wait for Processing

Step 11: Download Processed Results

Navigate to Output Folder

Understanding Output Structure

View Contents of a Result Folder

Download Results

Understanding the JSON Output

Useful SFTP Commands

Troubleshooting

Connection Issues

Upload Issues

No Output Files

Permission Errors

Best Practices

Security

File Management

Monitoring

Workflow Optimization

Next Steps

Additional Resources

Summary