Beyond the Scanner: A System for Organizing Your Digital Paperwork

You have a scanner, or maybe just your phone. The paper receipts, invoices, and forms that once cluttered your desk now clutter a folder on your computer. It feels like progress, but when you need to find that one expense from three months ago, you’re back to square one: opening dozens of files named scan_2024-08-17.jpg.

The problem isn’t digitizing the paper; it’s the lack of a system for the digital copies. A folder of images is just a digital pile. To be useful, that data needs to be extracted, structured, and findable. It needs a home.

The answer is to treat your scans not as individual files, but as entries in a database you control. In space-ocr, your workspace is a tree of folders you create — one per client, one per financial quarter, whatever matches the way you work. It’s your filing cabinet, and it holds more than one kind of thing: sheets, document bundles, memos, and the uploaded images themselves.

Inside a folder, you create a Sheet. It is a table whose columns you define once for one type of document. For your receipts, you might define columns for Store Name, Date, and Total. You only do this once. From then on, every receipt you drag into that sheet is automatically processed, and the extracted data lands as a new, organized row.

Not every document belongs in a table. A contract, a report, a multi-page notice — forcing those into columns loses the part you wanted to read. Those go into a document bundle instead: markdown mode when the headings and tables matter, text mode when you only need the words in reading order. Uploaded pages stack up inside the bundle, and a memo next to it holds the context the paper never carried.

The same tree is readable from code. GET /space walks it — give it a path and a depth, and you get back the items below with their types. GET /view reads one of them: a sheet comes back as rows with their values, coordinates and review result, a bundle as its pages. Sheet queries run on the server, so where filters, sort orders, select picks the columns, and limit with offset pages through the result instead of pulling the whole sheet down. Neither call spends a credit.

This turns a chaotic process into a routine. A hundred receipts are no longer a hundred files to manage; they are a hundred rows in a single, consistent table. You can even drop in a PDF with multiple pages; each page becomes its own row.

Of course, automated extraction needs to be trustworthy. If the system reads a total as 2,045, how do you know it’s correct? You click the cell. The tool instantly highlights the exact location on the original document image where it found that number. You can see for yourself. If you ever find a mistake, you can correct it by typing directly into the cell. Your manual entry will always override the OCR value.

This system is built for practical use, which extends to the pricing. It’s a simple pay-as-you-go model at $0.05 per scanned image. Your first 100 scans each month are free, and if an OCR scan fails for any reason, there is no charge. Browsing your folders and opening a sheet cost nothing, in the app and over the API alike. Exporting is where the plans differ: a CSV download spends one credit — the same unit as one scan — on the Free and pay-as-you-go tiers and is free on Starter and Pro, and downloading the same data again is never charged on any plan.

For more regular use, flat-fee plans provide more capacity. The Free plan includes 3 sheets and 1 GB of storage. The Starter plan raises that to up to 15 sheets, 10 GB of storage, and 500 scans a month (400 from the plan on top of the 100 free ones), while the Pro plan offers unlimited sheets, 100 GB of storage, and about 1,100 scans per month.