Frequently Asked Questions

🧠

Smart Prompt Engines & Dynamic Personas

Rusper doesn't just transcribe raw speech; it refines your spoken stream-of-consciousness through specialized LLM prompt personas. You can switch personas with one click in the Dashboard or create your own custom prompt:

  • •
    💻 Developer & Technical Specification: Preserves code signatures, camelCase, JSON formats, wraps variables in backticks (`userId`), and detects spoken length/expansion cues.
  • •
    🔥 Smart Self-Correction & Emotion Stripping (Banger): Erases verbal venting, frustration, and throat-clearing while seamlessly resolving mid-sentence plan revisions.
  • •
    ✉️ Professional Email & Workplace Message: Formats spoken rambles into clean, structured corporate emails or Slack updates with proper paragraphs, bullet points, and greetings.
  • •
    📝 Executive Summary & Action Items: Condenses meetings and brain dumps into scannable markdown bullet points.
  • •
    ✍️ Clean Verbatim: Minimal polish that fixes spelling, punctuation, and stutters while keeping 100% of your exact spoken words intact.
  • •
    ⚙️ Custom Persona: Write your own system instructions to format text however your specific workflow demands.
✂️

Voice Smart Editing

Voice Smart Editing is Rusper's superpower for rewriting existing text anywhere in Windows without manual copy-pasting:

  1. Highlight any text in any application (VS Code, Word, Chrome, Slack, Notion, etc.).
  2. Press your Smart Edit hotkey (default Ctrl + Alt + Space or custom key).
  3. Speak your instruction (e.g., "translate this to Spanish", "make this punchy", "turn into bullet points", "fix spelling and grammar").
  4. Release: Rusper sends your selection + voice instruction to Groq, transforms the text, and instantly replaces the highlighted selection in-place.
📚

Custom Vocabulary & App-Aware Context

⌨️

Global Hotkeys & Native Windows Injection

🎙️

Operating Modes & Glassmorphic HUD

⚡

Sub-300ms Speed, Privacy & $0 BYOK Model

Rusper runs on Groq's specialized Language Processing Units (LPUs). Groq LPUs execute OpenAI's Whisper Large v3 Turbo model in the cloud at ~216x real-time speed, giving you instantaneous transcription faster than local GPUs and without heating up your computer.