Working on a multi-tool web platform centered on an AI text humanizer. Stack is plain HTML/JS on the front, Firebase for auth, and the Anthropic Claude API on the backend.
Status
- Core humanizer flow works end to end
- Detection benchmarks in testing look strong so far
- Shared front-end shell designed for 14 planned tools
- Auth layer still being tightened before a wider rollout
What's next
Two threads I'm thinking about: finishing the tool roadmap on one shared codebase, and eventually migrating inference to a local LLM on a Jetson Nano so we're less API-dependent. Neither is done yet—this post is the checkpoint.