Downloading and running the SPARK OCR module

Receipt OCR runs on your own computer, not on the Nexus One server — that way it always uses your machine's own CPU cores, and your receipt images never leave your computer or get uploaded anywhere during the scan.

Download it

Click Batch OCR in SPARK's toolbar. It downloads SPARK_OCR.zip — unzip it anywhere. Inside you'll find SPARK_OCR.exe, an _internal folder, setup_tesseract.bat, and a README.

Keep SPARK_OCR.exe and _internal together. The _internal folder holds everything the exe needs to run. To put the tool somewhere else, move the whole folder (a desktop shortcut to the exe is fine) — copying just the exe out on its own will not start.

First run

Windows will likely show "Windows protected your PC" the first time you run the exe — it isn't digitally signed yet, not a sign anything is wrong. Click More info, then Run anyway. You should only see this once.

One-time setup

Double-click setup_tesseract.bat to install Tesseract-OCR, the OCR engine the exe relies on, via Windows Package Manager (accept the admin prompt if asked). Already installed? The script detects that and skips straight to done.

Running a scan

Open SPARK_OCR.exe, set a Receipts Folder and optionally an Output Folder, adjust Worker Processes (defaults to your CPU count minus one, up to 64), and click Run OCR Scan. Stop ends a run early; Open Results opens the Excel report once it's done.

Keep the companion transaction export — the file with a "Transaction - Receipt Image Reference ID" column — in the same Receipts Folder. That's how the scan links a flagged receipt back to a real Transaction Reference; without it, OCR still runs, but its output won't carry a Transaction Reference or Image Folder Path.

Bringing results into SPARK

Click OCR Excel in SPARK's toolbar and load the Excel file the scan wrote.