ToneShift is a lightweight, AI-driven text transformation engine. It processes raw input and instantly restructures it into specific stylistic formats—whether you need business-ready polish, authentic conversational pacing, or structured prompts for other AI agents.
ToneShift uses LLaMA 3.3 (via Groq) to analyze and rewrite text across five distinct modes:
| Mode | Function |
|---|---|
| Professional | Polishes rough drafts into business-ready copy suitable for executive communication and formal documentation. |
| Casual | Relaxes formal text into natural, conversational language with authentic phrasing. |
| Humanize | Strips predictable LLM cadence and vocabulary, restructuring text to bypass synthetic detection. |
| AI Prompt | Expands raw ideas into highly structured, context-rich prompts designed for other AI agents. |
| Detector | Analyzes text patterns to estimate the probability of synthetic generation. |
ToneShift requires a free Groq API key to handle the model inference.
ToneShift can run entirely in the browser without a build step.
- Clone the repository.
- Duplicate
.env.exampleand rename it to.env.local. - Add your Groq API key to the file.
- Open
toneshift.htmldirectly in your browser.
For production use, the backend is optimized for Vercel Serverless Functions.
- Install the Vercel CLI:
npm i -g vercel - Authenticate:
vercel login - Deploy the project:
vercel - In your Vercel Project Settings, add a new Environment Variable:
GROQ_API_KEYmapped to your key.
ToneShift is built with a minimalist, dependency-light stack:
- Frontend: Vanilla HTML/CSS/JavaScript (Clean, dark-mode-ready UI)
- Backend: Node.js (Vercel Serverless Functions)
- Inference: LLaMA 3.3 powered by Groq's high-speed API
ToneShift's transformation logic can be consumed by external applications via its serverless endpoint.
Endpoint: POST /api/transform
Payload:
{
"text": "String to be transformed",
"tone": "professional | casual | humanize | prompt | detector"
}