April 2025: searching inside your documents, not just their names
A library is only as good as your ability to find things in it. In April 2025 we made the library search inside the files, not just at their names.
Part 21 of 39 of Building TradeYoke, the story of how the platform was built, from January 2024 to today. Previous: April 2025: what happens after a quote is sent. Next: April 2025: expiry dates, so a certificate never runs out unnoticed. See every part on the build timeline.
Quick answer
Quick answer: TradeYoke's document search reads the text inside PDFs, Word, Excel and PowerPoint files, plain text, web pages and scanned images (using optical character recognition), and shows a snippet of the match. A search for a serial number or a customer's name finds the document that mentions it, whatever the file is called.
Key takeaways
- Search looks at the words inside the file, not only at its name.
- It reads PDFs, Word, Excel and PowerPoint, text, HTML and scanned images.
- Results show a snippet, so you can tell the right one at a glance.
- New and older documents are indexed in the background, so nothing needs re-uploading.
The file called scan0042
Every business has one. A scanner or a phone names the file for you, and months later you are looking for the certificate that mentioned a particular boiler, and all you have is a list of names that mean nothing. Filing against the job and customer helps, but only when you remember which job it was.
What the search reads
In April 2025 we added a full-text index to the library. It reads the text inside the file:
- PDFs, including the text layer.
- Word, Excel and PowerPoint files.
- Plain text and web pages.
- Images and scans, using optical character recognition, so a photographed page can be found by the words printed on it.
A search then shows a snippet around the match, so you can often tell it is the right document without opening it.
Indexing that looks after itself
An index that has to be rebuilt by hand is an index nobody trusts. New documents are read as they arrive, and a background pass works through anything that was uploaded before the index existed, so older paperwork becomes searchable without anyone re-uploading it. If a file cannot be read the first time, it is picked up again on a later pass.
Video too
A short video taken on site can hold the one useful sentence, such as a customer describing what is wrong. Videos can be transcribed, and the transcript is searchable alongside the documents.
What it is good for
- Finding the certificate that mentions a serial number.
- Finding every document that names a particular customer or street.
- Finding the manual page about an error code (the Knowledge shelf uses the same search).
Next
Later in April: dates. A certificate you can find is good; one you are told about before it runs out is better.
Questions people ask
Can TradeYoke search inside a scanned PDF or a photo of a page?
Yes. Scanned pages and images are read with optical character recognition, so a search for words printed on them finds the document.
Does searching documents work on older files I uploaded earlier?
Yes. A background pass indexes documents uploaded before search was added, so you do not need to upload them again.
Try it on your own week
30 days of Essentials free, no card. Bring your customers and see it against a real week.
