Invisible · Flexible · Powerful

Your silent
interview co-pilot

Interview Assistant captures the problem on screen, sends it to any AI service you configure, and overlays the solution — completely invisible to screen-sharing software.

Download for Windows See how it works ↓

Free to use · Bring your own API key · No telemetry

Interview Assistant

You · screenshot.png

Analyze the problem in the screenshot and solve it in C++.

Gemini 2.5 Flash

Thinking…

// Two Sum — O(n) hash map
vector<int> twoSum(vector<int>& nums, int t) {
  unordered_map<int,int> seen;
  for (int i = 0; i < nums.size(); i++) {
    int comp = t - nums[i];
    if (seen.count(comp))
      return {seen[comp], i};
    seen[nums[i]] = i;
  }
  return {};
}

Core capabilities

Everything you need,
nothing you don't

👁

Completely invisible

Undetectable by Zoom, Teams, Meet, and all screen-sharing tools. Hidden from the task manager and browser APIs.

⚡

Instant hotkeys

One-key screenshot, send, and receive. Full keyboard control — no clicks, no mouse, no suspicious pauses.

🔧

Any AI service

Connect Gemini, Claude, GPT, Ollama, Yandex AI, or your own endpoint. Define custom request templates in YAML.

💬

Multi-turn chat

Build on previous answers. Ask follow-ups, refine solutions, dig deeper — with full conversation history.

🖼

Screenshot & crop

Capture the full screen or draw a precise crop. Attach images, PDFs, or any files alongside your prompt.

🌐

Proxy support

Route requests through any proxy to bypass regional API restrictions — configurable per service.

Workflow

Four steps to the offer

Launch before the interview

Start IA — it hides itself from all screen-sharing software. Invisible to the interviewer.

Interviewer shows the problem

A coding challenge appears on screen. Press Ctrl+S to capture it instantly.

AI analyzes & responds

The screenshot is sent to your configured AI service. A full solution with explanation appears in the overlay.

Win the offer

Read the solution, understand the reasoning, implement it. Confidently.

Hotkeys

Full control,
zero visibility

Every action is a keystroke. No clicks to reveal, no window to spot. All hotkeys are remappable in settings.toml.

Ctrl+STake screenshot

Ctrl+OCrop & select area

Ctrl+EnterSend request

Ctrl+HToggle visibility

Ctrl+KSwitch page/chat mode

Ctrl+QQuit

Flexibility

Connect any AI service

Describe any HTTP endpoint in services.yaml. Use Inja-powered templates with variables, loops, and conditionals. Switch between providers in one click.

Google Gemini

Yandex AI

Ollama (local)

Anthropic

OpenAI

Your own API

services.yamlYAML

serviceProviders:
  - name: Google
    services:
      - name: Gemini/gemini-2.5-flash
        baseUrl: "https://generativelanguage.googleapis.com/..."
        extends: GoogleServiceBase
        requests:
          - name: AnalyzeTaskImageEng
            extends: RequestBase
            display:
              reply:
                selector: "{{ get(parse_json(ServiceReply.Body), \"candidates.0.content.parts.0.text\") }}"

  - name: Ollama
    services:
      - name: gemma3:27b
        baseUrl: "http://localhost:11434/api/generate"
        requests:
          - name: AnalyzeImages
            body: |
              { "model": "gemma3:27b", "prompt": "{{ default(CustomPrompt, RUS_PROMPT) }}" }

templates: reusable request bases — extend and override any field.

globals: shared variables like API keys, rendered with Inja.

display selectors: JSONPath-style extraction from any response body.