Skip to guide content

Guide

How Aspera works

Aspera is a stealth AI coding interview overlay: hide the app, capture what is on screen, analyze with your model of choice, and deliver solutions where only you can see them.

The product loop

  1. Hide / stealth — Keep the overlay invisible during shares and presentations. No floating UI, no notifications, no trace on the shared surface.
  2. Capture — Grab any region of the screen with a global hotkey. OCR extracts code and text from what you selected.
  3. Analyze — Send the capture to Gemini, ChatGPT, or Claude with your own encrypted keys. Use Coding mode for algorithms or General mode for wider problems.
  4. Deliver off-device — Read results in private chat when appropriate, or push them to email or Telegram so answers live on another device — not on the shared screen.

What that enables

  • Coding interviews — Capture a problem or editor, get a walkthrough and fixed solution path without leaving a visible assistant window in the share.
  • Stealth screen assistance — Stay in the conversation while analysis and follow-ups happen off the shared display.
  • Audio when you need it — Record system audio or the mic; Whisper transcription can feed the same AI path, with optional Telegram delivery.

How this differs from floating-overlay tools

Many tools assume an always-visible or “invisible floating” overlay inside the shared desktop. That model is brittle when screen share captures everything, and it is often weak or missing on Linux. Aspera leans on hide the whole app, hotkey capture, and answers on another device — a loop that still works when classic overlay stealth fails. More on the Linux wedge: Built for Linux.

Your keys stay yours

Bring your own API keys. They are encrypted in the OS keychain, stored locally, and not sent to Aspera servers for model calls.