AlgoVortex

Local AI coding-interview practice. Desktop-first, private by default, and powered by Ollama or Gemini.

Setup | Running | Ollama Setup | AI Provider | Voice Setup | Build | Contributing

AlgoVortex is a fully local desktop app for practicing coding interviews with an AI interviewer. Progress and settings stay in local storage on your device — no cloud accounts required.

Download

algovortex.net — visit the landing page to download the latest Mac release.

Currently only available on macOS (Apple Silicon).

Use AlgoVortex with Ollama for local models, or paste a Gemini API key if you prefer a hosted AI provider. The Electron app gives the smoothest local Ollama experience because it can talk to your local server without browser CORS issues.

Prerequisites

• Node.js 18+
• One AI backend:
  • Ollama (recommended): you must install Ollama separately, then pull a model. See docs/ollama-setup.md for full steps.
  • Gemini: a Gemini API key from aistudio.google.com/app/apikey

Note: AlgoVortex does not include Ollama. The Electron app can try to run ollama serve on launch if the server is not already running, but only after Ollama is installed on your machine.

Setup

npm install

If you plan to use local AI, install Ollama and pull a model before running the app. Quick start:

# After installing Ollama from https://ollama.com/download
ollama pull llama3.2

See docs/ollama-setup.md for macOS, Windows, and Linux install instructions.

Running

Desktop app (Electron, recommended)

npm run electron:dev

This starts the React dev server and opens the Electron window automatically.

Browser only (React dev server)

npm start

Then open http://localhost:3000.

Note: when running in a normal browser, calls to a local Ollama server may be blocked by CORS. The Electron app disables this restriction for local content, so the desktop app is the smoothest experience for Ollama.

Configuring the AI provider

1. Launch the app and open Settings in the navbar.
2. Choose Ollama (Local) or Gemini (API Key).
   • Ollama: install Ollama first (setup guide), set the base URL (default http://localhost:11434) and model name (e.g. llama3.2). Use Test Connection to list installed models.
   • Gemini: paste your API key and (optionally) change the model.
3. Optionally configure Voice / ElevenLabs — see below.
4. Click Save Settings.

Voice setup

Voice is optional. AlgoVortex can speak interviewer responses in two ways:

• Built-in web/OS voice: free and enabled automatically when no ElevenLabs key is saved, or when ElevenLabs is unavailable or out of credits.
• ElevenLabs voice: optional API-backed voice. A paid ElevenLabs tier gives more realistic, natural-sounding speech and higher usage limits than the built-in web voice.

To use ElevenLabs:

1. Create an account at elevenlabs.io (the free tier can be used to test API access, but usage and voice quality/limits may vary).
2. Go to elevenlabs.io/app/settings/api-keys (Profile → API Keys) and create a new key.
3. Copy the key.
4. In AlgoVortex, open Settings → Voice (ElevenLabs) and paste the key into the ElevenLabs API Key field.
5. Click Save Settings.

When no key is set, the interviewer still speaks using the free built-in web/OS voice. If an ElevenLabs key is invalid, out of credits, or blocked by the account tier, AlgoVortex falls back to that built-in voice so the interview can continue.

Voice input/transcription uses ElevenLabs Speech-to-Text, so speaking answers into the microphone requires an ElevenLabs key with Speech-to-Text access. Typed answers and spoken interviewer replies still work without that key.

Contributing

Issues, pull requests, bug reports, and docs fixes are welcome. Read CONTRIBUTING.md before opening a pull request.

License

AlgoVortex is available under the MIT License. See licenses/LICENSE-MIT.md.

Copyright (c) 2026 Francisco Figueroa.