Inspiration

We wanted a tool that turns simple intentions into beautifully phrased messages instantly — like browsing “genres” on Netflix, but for conversations.

What it does

It takes one short input (like “I missed you”) and generates multiple phrasing options in different tones such as Formal, Cheesy, In-Love, or Respectful.

How we built it

We used Google GenAI for generation, sentence-transformers + ChromaDB for tone-based retrieval, FastAPI for the backend, and a warm writer-themed React UI.

Challenges we ran into

Managing consistent tone across generations, embedding thousands of phrasing variations efficiently, and designing an intuitive genre-based UX.

Accomplishments that we're proud of

We turned a simple idea into a fast, aesthetic, tone-aware phrasing engine that feels fun and extremely easy to use.

What we learned

How to mix LLM generation with vector search, build clean React storytelling UIs, and structure prompts for consistent conversational tone.

What's next for tone-weaver

More tones, personalization profiles, a browser extension, and real-time “rewrite my message” suggestions across apps.

Built With

  • chromadb-line-2:-python
  • docker
  • fastapi
  • github
  • gitpython
  • google-genai
  • python-dotenv
  • react
  • sentence-transformers
  • tailwind-css
Share this project:

Updates