I’m Building 100 Tools With AI

Tags: AI | Indie Development | Vibcoding | Open Source

I used to be a backend engineer. Building APIs, setting up services, tuning queues and keeping systems stable. Every project started with careful tradeoffs around architecture, standards and maintainability. Solid, but heavy. Turning an idea into reality used to require a long planning journey.

Now I’ve switched up my workflow: vibcoding. No longer chained to rigid engineering rituals. I don’t wait for everything to be perfect before starting. When an idea pops up, I use AI as my lever and rapidly build a working prototype.

I’ve set myself a wild, fun goal: build 100 tools with AI.

These won’t be polished SaaS products, venture-backed pitches or massive platforms for millions of users. They are small, sharp utilities, each built to solve one tiny, real, annoying pain point: local PDF redaction, batch format conversion, browser automation scripts, lightweight RAG assistants, log organizers, batch processing scripts, data validators, document converters… Some are built for developers, some fix office headaches, and many exist just to solve my own needs.

One hundred sounds like a big number. Break it down, and it becomes approachable: One pain point, one implementation, one release — that counts as one tool.

My years of backend experience are my foundation. I understand concurrency, storage, edge cases and security boundaries. AI is my new multiplier. It shrinks the timeline from idea to working version dramatically. What once took days to prototype can now be validated in hours.

Still, I hold one clear rule: Speed does not mean sloppy work. Tools handling files or sensitive data like redaction run locally whenever possible; your original files never leave your device. Input validation and safety guards stay in place where they matter.

These 100 tools will vary a lot: Some will prove useful and get ongoing iterations. Some are just experimental demos, archived once finished. Others will turn out to be useless once built. And that’s okay. Failed prototypes are part of vibcoding.

All tools will be open-sourced here, with ongoing progress updates. Feel free to try them out, explore and open issues: 👉 https://www.qudo.cc/en/open-source-tools

No overblown promises, no guarantees of perfection. Just keep building.

This is my long-term experiment: A backend engineer, embracing the AI era, forging small tools one after another.

Target: 100. The journey starts now.

Comments

Post a Comment

Popular posts from this blog

Why I built a free, offline PDF tool.

Image
  Fed up with PDF software constantly pushing "Premium Memberships"? So was I! Need to add a watermark? VIP. Want to sign a document? VIP. Even just converting a format? VIP. Since nobody was making a clean, straightforward tool, I decided to build one myself. Check out the screenshot below—this is the constant "Upgrade to Premium" nagging I got while using a certain popular PDF software: With the rise of AI, I realized that core engine code I wouldn't have dared to write before is totally doable now—as long as you have enough Tokens. After getting completely annoyed by all the paywalls, I decided to hand-code this myself. I did build an online version first, but I know many people worry about security when processing files online. Even though my online version processes everything locally in the browser using JS (no files are ever uploaded to the cloud), I figured an offline app would be even better. So, in my spare time, I let AI help me with the vibe coding. ...
Read more

[Open Source] Qudo PDF Toolbox - Offline, Fully Local Desktop PDF Toolkit

Image
  Project Introduction https://github.com/xuejiazhi-qudo/qudo-pdf-tools   Qudo PDF Tools (qudo-pdf-tools) is an offline, fully local desktop PDF toolkit. It runs as a standalone Windows desktop application, with all processing performed on the local machine—no server-side components, no internet connection, and no telemetry data collection. Files never leave the device from start to finish. It is more than just a format-conversion utility; it is a desktop application capable of genuine PDF editing. Inserted text and images are floating elements that can be dragged, resized, and double-clicked to modify their content. Existing text, text blocks, and images on the page can also be directly clicked to edit text or dragged to new positions. The interface supports real-time switching between Chinese and English. Open Source License AGPL-3.0 License (dependent components include PyMuPDF and pdf2docx. Free for personal and open-source use without restrictions; for closed-source comme...
Read more