Start practicing
Choose a template or describe your own interview, then practice with voice or text.
Documentation
A good place to begin.
From your first practice to your first contribution. Find the short guide, then go deeper when you need to.
Run and contribute
Build your own copy and help make interview practice useful for more people.
Understand the system
Read the design, data contracts and operational checks behind the application.
Community and privacy
Help test the beta, understand your controls, or contact the maintainer.
Models, keys and voice
Choose the model for your practice.
Hosted or personal
The free hosted interview uses Gemini 3.8 Flash for text and feedback. Each account gets one free interview; the project funds up to 200 interviews per UTC day. Personal-key setup supports Gemini, OpenAI, Anthropic, DeepSeek, xAI and Meta. Available models depend on your provider account; validate your selection before starting.
Your key is sent securely to the API, encrypted for the active interview, and expires. It is not saved in browser storage. Validation can make a small billable request.
Know what powers the session
Native live voice uses Gemini’s audio model. Other providers support text interviews and feedback; selecting one does not switch the voice engine. Use text when you want practice entirely on your selected provider.
Your selected provider and model stay attached to the attempt and its feedback. Provider limits, model availability and charges still apply when daily app limits do not.