My first published AI built project. Built to help me practice French.
Coded with Claude
Published via Netlify
I previously tried to build an app on Replit in 2025.
I failed to get it to a point where it could be published. It was too buggy and broken and I think I just didn't have enough technical knowledge to steer AI in the right direction. Lesson learned was really the importance of domain expertise, scoping and planning. PRDs (Product Requirement Documents) etc.. are still important
I think coding, even if it is now more accessible to a lot more people via AI and LLMs it seems it is still important to have a solid foundation to really be able to create effectively. I would liking it to film creation in someway or just content creation in general the idea that you need to know the rules and then break them.
For you to really be able to do powerful things with AI I still think that some foundational knowledge of the domain of any domain is important so learning and really diving into a crafter field is still super important. I think what AI does is elevate the baseline; there is no or there is little room to just do. You have to really be about your craft. You have to go deep.
Claude — All code generation, content creation, debugging
Netlify — Hosting + serverless functions (free tier)
ElevenLabs — Text-to-speech (Charlotte voice, free tier)
Google Gemini Flash — AI-based translation scoring (free tier)
Web Speech API — Browser TTS fallback
[In Progress] My first failed attempt - revisited
Coded with Claude
Published via Netlify
Github repo
First project gave me the confidence boost to try again with proskills (failed Replit project). Surprisingly using a lot of what I learned from the first project about no/low code software basics - how to deploy, using repos etc. Claude is also a lot more comfortable of an environment to tinker. Considered using Claude code but I think I have to in