As a modern-day knowledge worker, I have several thousand documents, presentations and other files on my computer. I built Buzee in my free time to help me weave my way through this maze. I have been using it pretty much everyday since the day I built it - and I love it!
I thought I could turn Buzee into a startup. I reached out to offices and helped set it up for them. But it didn't pan out.
I am now letting go of this project because I have other priorities in life.
Please feel free to do with this project as you wish. I am happy to help you get started with the codebase.
Do share what you build. I would love to see it!
Cheers
Also this is not the first time I'm seeing a universal search startups fail, some even YC backed. It sounds like an interesting idea, but why is it failing, I can't seem to figure out, maybe not enough market or adaptation?
The original idea emerged from my workplace where everyone has thousands of documents on their machines but can't recall what's where, so they always ask each other for files, and then end up with several duplicates. It's a mess. I thought a natural language powered high-speed, low-footprint app would be a great USP, especially when you compare it against Windows Search.
Turned out, everyday office people (not nerdy hackers) have their own set ways of managing and discovering files on their machines. A guy I met keeps ONE 1000-page Word document instead of hundreds of smaller Word files because Ctrl+F in a Word document is more effective than Windows Search! Plus, a lot of files are now moving online to Google Drive or are locked behind Microsoft OneDrive, especially if it's a corporate workplace. Corporate work machines also made it quite impossible to install an app with that isn't code signed. So my target audience was kind of priced out for me!
I didn't start out to make an 'universal search app' but I fell for its allure somewhere along the way. I wanted to make a simple desktop search app that _just works_. Kind of like Everything [1] for Windows but full-text, cross-platform and less nerdy.
In the end, what I got was an app that fits my needs perfectly and I am really happy with it!
Once I had a more reliable foundation, I planned on adding a tiny LLM that can run locally and parse user’s natural queries in a manner that the app can understand.
I didn’t really have a workflow to do anything since I am not a professional developer. I would just hack around in a dev branch, merge it with main and create a build when I was happy with it.
Recoll will index an MS-Word document stored as an attachment to an e-mail message inside a Thunderbird folder archived in a Zip file (and more… ). It will also help you search for it with a friendly and powerful interface, and let you open a copy of a PDF at the right page with two clicks. There is little that will remain hidden on your disk.
For Windows, you should try this link [1]. Unfortunately, it'll only work if you install it under admin privileges since I couldn't afford to buy a $500 code signing certificate. Otherwise, you could build the app on your machine from the source code! I'd be happy to help you with that.
Let me know how it works out.
https://github.com/gerardog/gsudo/issues/1#issuecomment-2111...
suggests either:
https://github.com/Azure/trusted-signing-action
Which is 10 USD/month - or publishing in the windows store:
https://learn.microsoft.com/en-us/windows/apps/publish/?tabs...
https://github.com/marketplace/actions/microsoft-store-publi...
Not free either, but one time fee:
But yes, it absolutely sucks and it’s very easy to lose hope when developing for desktop. Every platform sucks in unique ways.
Hit me up if you wanna chat. /A fellow tauri app developer
Absolutely bonkers, as the only metric is creation time, so good luck if you have a precious folder you forgot, with old contracts or something like that.
Here's mine in her thieving puppy phase. https://imgur.com/a/wpJD4SI
You can see Buzo’s mugshot here: https://buzo.tools/
:D
And there's a lot of enterprises that use that and pay a lot of money for it! The business side problem i can see is : how do you convince MS focused companies to use your product instead of SharePoint? Could your product be built on top of it?
Can you for example also enhance gSuite?
Thank you for open sourcing this, so others may learn from you or build upon your work!
I have updated the readme with screenshots. You can visit the website [1] or the releases page [2] to try it out.
[1]: https://buzee.co/
https://duckdb.org/community_extensions/extensions/hostfs.ht...
I ask because in my own project I have my own search query grammar [0] written in Lezer that more or less "transpiles" [1] to SQL+FTS5 that's been uh, tedious to build.
[0] https://github.com/AlexErrant/Pentive/blob/main/shared-dom/src/language/query.grammar
[1] https://github.com/AlexErrant/Pentive/blob/main/shared-dom/src/language/query2sql.ts(For your next project, may I suggest double checking if the name overlaps with any colorful slang?)
btw had a request, I saw your repo lacks proper documentation. We are building an open source ai documentation system, I would love to try if I may have your permission to build docs and create a PR!
You can checkout our project here: https://git.new/akiradocs
My intention was to make it free for individuals (search on desktop + personal cloud) but charge business users who would want shared access on a network. Any AI features would also be free if it runs locally or chargeable for Claude/GPT. I wanted to leverage Mac and Windows’ built-in AI/ML capabilities as much as possible. The OCR on both systems is actually really powerful! Even my ten year old Asus laptop was able to OCR PDFs and images very quickly and reliably.
Local and privacy first was my core principle.
In case it doesn't work still, I'd suggest trying to build from source. I wouldn't be surprised if your machine is an edge case for Windows that I missed.
The main motivation was that Windows Search is often slow and terrible. But then I wasn’t able to help many Windows users because of corporate security on their work machines and my lack of $500 Windows code signing certificates.
The main sell was that it was fast and did full text search using natural language.
I had several other plans but wasn’t really able to make much of it happen. In the end this became a passion project that I tweaked endlessly to learn new technologies rather than create value for the user.
I was trying to scratch my own itch really.
Still, I like Buzee's UX and 10/10 agree it would be a major step up for Windows users.
That was part of my failure. I wanted to solve a problem for Windows but didn't have a decent Windows machine myself! Had to rely on a 10-year old laptop to do testing and building, which inevitably didn't work well.
Will add an architectural diagram to the readme and in-line comments to the code.
At refind.ai (https://refind.ai/), we’ve been working on a similar challenge of navigating large volumes of documents, but with a more AI-driven and user centric approach. Beyond just search, we focus on transforming unstructured data into structured insights. This includes features like automatic metadata extraction, natural language search, and integrations with external systems like email and CRMs.
To those who struggle with file management at scale, especially in environments where tools like Spotlight or Windows Search fall short, I think there’s a lot of potential for tools like Buzee to evolve further. If anyone wants to collaborate or learn more about our approach, feel free to reach out!
Best of luck with Buzee’s journey—I’d love to see how the community builds on it!