⚡ Promptini

Turn a rough idea in Tunisian Derja, Arabic, French, or English into a clean, model-ready prompt, without leaving VS Code.

Promptini reads messy, code-switched developer input and rewrites it into a precise, structured prompt you can paste straight into Claude, ChatGPT, Gemini, or Cursor. Whatever language you type in, the optimized prompt always comes out in clear English.

Features

Derja-first understanding — handles authentic Tunisian Derja and Arabizi transliteration (3 → ع, 7 → ح, 9 → ق), mixed freely with Arabic, French, and English.
Model-aware formatting — structures the prompt for your target tool: Claude, ChatGPT, Gemini, Cursor, or a generic "Any".
Always-English output — type in any language; the result is a clean English prompt.
Sidebar panel — a dedicated Activity Bar view: type an idea, pick a model, get the optimized prompt in one click, and copy it.
Optimize Selection — highlight rough text anywhere in the editor and turn it into a structured prompt from the right-click menu.

Getting started

Click the ⚡ Promptini icon in the Activity Bar to open the panel.
Type or paste a rough idea and choose your target Model.
Press ⚡ Promptini, then Copy prompt.

By default the extension uses the hosted Promptini engine, so there is nothing to install and no API key to configure. To run the engine yourself, see Cloud vs. local.

Optimize a selection

Select text in any editor, right-click, and choose:

Promptini: Optimize Selection — sends the selection to the sidebar panel.
Promptini: Optimize Selection via API — optimizes it through your local engine and replaces the selection in place.

Commands

Command	Description
`Promptini: Open App`	Open the full-page Promptini panel in the editor.
`Promptini: Optimize Selection`	Send the highlighted text to the sidebar.
`Promptini: Optimize Selection via API`	Optimize the selection through the local API and replace it in place.
`Promptini: Stop App`	Stop the local engine and close the panel.

Open the Command Palette (Ctrl/Cmd + Shift + P) and type "Promptini" to find them.

Extension settings

Setting	Default	Description
`promptini.mode`	`cloud`	`cloud` uses the hosted engine; `local` runs your own Python server.
`promptini.hfUrl`	hosted Space URL	URL of the hosted Promptini engine (used in cloud mode).
`promptini.apiUrl`	`http://127.0.0.1:8000`	Base URL of your local engine, used by Optimize Selection via API.
`promptini.apiModel`	`general`	Target model used by Optimize Selection via API.
`promptini.pythonPath`	`""`	Optional path to a Python interpreter for local mode. A workspace `.venv` is auto-detected otherwise.

Cloud vs. local

Cloud (default). The extension calls the hosted Promptini engine. Nothing to install; it works out of the box.
Local. Set promptini.mode to local to run the engine on your own machine. The extension starts the FastAPI app (uvicorn api:app) from your workspace, so you need the Promptini repository open as the workspace folder with its Python dependencies installed.

Requirements

VS Code 1.88.0 or newer.
For local mode only: Python 3.11+, with the Promptini repository open as your workspace and its dependencies installed (pip install -r requirements.txt).

How it works

Promptini sends your input to a prompt-engineering engine that detects the task type, understands Derja and Arabizi natively, and rewrites the request into a model-specific structure — for example # ROLE, # TASK, # REQUIREMENTS, # OUTPUT FORMAT, and # CONSTRAINTS for Claude. It runs on a fully free, open-source stack: Groq for inference and Hugging Face for hosting.

Known issues

Local mode requires the Promptini repository open as the workspace folder; it will not start without api.py present.
The hosted engine may take a few seconds to wake on the first request after a period of inactivity.

Promptini AI Prompt Studio

Becher Zribi