Skip to content

quickstart.md

Latest commit

 

History

History
177 lines (115 loc) · 5.57 KB

quickstart.md

File metadata and controls

177 lines (115 loc) · 5.57 KB

Quick Start Guide

Get Promptimize running in a few minutes.

What Promptimize Does

Promptimize has two separate services and two recording modes:

  1. Voice-to-text (required) — Always uses OpenAI Whisper. Requires an OpenAI API key.
  2. Prompt optimization (optional) — Converts transcribed speech into structured prompts. Choose from 8 providers.
Mode Shortcut Pipeline
Transcribe Cmd/Ctrl+Alt+V Record → Whisper → insert raw text
Promptimize Cmd/Ctrl+Alt+P Record → Whisper → optimize → insert

See Recording Modes for when to use each mode.

graph LR
    Voice[Your Voice] --> Whisper[OpenAI Whisper<br/>Transcription]
    Whisper --> RawText[Raw Text]
    RawText --> Choice{Mode?}
    Choice -->|Transcribe| Editor[Insert raw text]
    Choice -->|Promptimize| Provider[Optimization Provider]
    Provider --> OptimizedText[Optimized Prompt]
    OptimizedText --> Editor
Loading

Installation

From VSIX (current)

  1. Download the latest .vsix from Releases
  2. Open VSCode or Cursor
  3. Extensions → ... menu → Install from VSIX...
  4. Select the downloaded file
  5. Reload the window if prompted

From Marketplace (coming soon)

Search for Promptimize in the Extensions view.

First-Time Setup

On first launch, Promptimize prompts you to open the Configuration panel. You can also open it anytime:

Command Palette → Promptimize: Open Configuration or click $(gear) Settings in the status bar.

Configuration panel overview

The panel is a single webview (not a multi-step wizard) with these sections:

  1. Transcription — Enter OpenAI API key, test connection
  2. Prompt Optimization — Enable/disable, choose provider, enter credentials, pick model
  3. System Prompt — Customize transformation style (optional)
  4. Provider Comparison — Compare all 8 providers
  5. Save & Close — Validates and completes setup

Full details: Configuration Webview Guide

Minimum configuration (transcription only)

  1. Open the configuration panel
  2. Enter your OpenAI API key → Test (optional)
  3. Leave Enable prompt optimization unchecked
  4. Click Save & Close

First Recording

Transcribe (raw text)

  1. Open an editor or Cursor chat input
  2. Press Cmd+Alt+V (macOS) or Ctrl+Alt+V (Windows/Linux)
  3. Speak clearly
  4. Click Recording... in the status bar to stop
  5. Raw transcription appears in your editor or chat

Promptimize (optimized prompt)

  1. Ensure optimization is enabled in the configuration panel
  2. Press Cmd+Alt+P (macOS) or Ctrl+Alt+P (Windows/Linux)
  3. Speak clearly
  4. Click Recording... in the status bar to stop
  5. Optimized prompt appears after Whisper + transformation

Note: Keyboard shortcuts start recording only. Stop via the status bar or stop commands.

Configuration Commands

Command Purpose
Promptimize: Open Configuration Configuration webview (primary)
Promptimize: Setup Wizard Same as Open Configuration
Promptimize: Configure OpenAI API Key (Whisper) Set or update OpenAI key
Promptimize: Configure Prompt Optimization Provider Command Palette provider wizard
Promptimize: Configure OpenAI Optimization Model Pick GPT model (OpenAI provider)
Promptimize: Test Configuration Test setup; opens before/after webview

Status Bar

Three items in the status bar (right side):

Indicator Meaning
$(mic) Transcribe Start/stop raw transcription mode
$(sparkle) Promptimize Start/stop optimized prompt mode
$(gear) Settings / $(warning) Setup Open configuration panel

While recording, the active mode shows $(record) Recording... (click to stop).

During processing, notifications show: Transcribing... → Optimizing... → Inserting...

Tooltip when idle: Transcription: OpenAI Whisper | Optimization: [Provider]

Test Configuration

Run Promptimize: Test Configuration to validate your setup. When optimization is enabled, a webview opens showing:

  • Original sample transcription
  • Transformed prompt from your provider
  • Improvements list (filler removal, conciseness, structure)

See Advanced Settings — Test Configuration.

Troubleshooting

See the full Troubleshooting Guide with decision trees.

OpenAI API key errors

Optimization provider errors

  • Each provider needs its own API key (except Ollama and OpenCode)
  • OpenAI for Whisper and OpenAI for optimization can use the same key
  • Reconfigure via Open Configuration panel

Microphone not working

macOS: System Settings → Privacy & Security → Microphone → enable Cursor/VSCode

Windows: Settings → Privacy → Microphone → enable Cursor/VSCode

Text not inserting

  • Focus an editor or chat input before recording
  • Check status bar for errors
  • Text may fall back to clipboard — paste manually

Need more help?

Next: Configuration Guide · Recording Modes