# OCR Engine

When setting up a new optical recognition program you need to enter the Config details before adding the details.

### General

1. Start by entering a Configuration ID.
2. Add a short description to be used to explain what the OCR is or for.

<figure><img src="https://content.gitbook.com/content/iyzfmS8DQ2bAtaOk1MxQ/blobs/KO0BXN9X2zPnjTizAVJL/image.png" alt=""><figcaption></figcaption></figure>

### Ocr Engine

1. Select the type from the Ocr Engine Type dropdown.
2. You can limit the characters recognised by entering them into Allow List Characters.
   * You must include a space-character to retain whitespace.
   * If left blank all characters will be included.
3. You can limit the characters recognised by entering them any into the Deny List Characters.
   * If left blank all characters will be included.
4. Choose the Language Mode from the dropdown.
   * English - The Default mode, it uses both an LSTM and OEM strategy to produce fast and accurate results.
   * English (Best) - Uses an LSTM engine optimised for detail over speed.
   * English (Fast) - Uses an LSTM tesseract engine optimised for speed over accuracy.

<figure><img src="https://content.gitbook.com/content/iyzfmS8DQ2bAtaOk1MxQ/blobs/2zeK6u3AfYLQVhhAlCb6/image.png" alt=""><figcaption></figcaption></figure>

### Agent

1. Select the Source Agent you want to use for this Source Configuration.

<figure><img src="https://content.gitbook.com/content/iyzfmS8DQ2bAtaOk1MxQ/blobs/nWfqCtHtjtrIoolhDpV0/image.png" alt=""><figcaption></figcaption></figure>