ANSLY
ANSLY
An AI assistant for technical interviews that hears the question, sees the screen and code, keeps context, and helps structure a precise answer.
What the business needed?
Ansly is a desktop product built for preparation and real-time support during technical interviews. It combines the interviewer voice, on-screen content, code, and follow-up history, so each suggestion is grounded in the complete conversation. We designed a coherent journey from product discovery and model selection to app download, personal API key setup, and a live session. The interface stays compact and calm: it helps organize the answer without competing with the interview itself.
What we built
Multiple context sources
A question can be spoken, refer to code on screen, and continue an earlier topic. The product must merge those signals into one session.
Real-time response pressure
During an interview there is no time to switch tabs, retype a task, or repeatedly explain the same context to an AI model.
Different models and workflows
Users choose OpenAI, Gemini, Claude, or Qwen and expect a consistent setup rather than provider-specific friction.
Trust in a desktop product
Screen, microphone, and local API key handling had to be explained clearly without overwhelming the first launch.
Complex capability, quiet interface
The system processes rich context while the user should only see the current question, answer structure, and session state.
One path from landing to session
The marketing site, macOS and Windows downloads, authentication, documentation, and app needed to feel like one product.
How the project was built
01
Designed a contextual session model
Voice, screen, code, and follow-up history are combined so every answer is based on the full interview context.
02
Built a compact live workspace
The interface focuses on the question, answer structure, and listening controls without distracting secondary panels.
03
Added multi-provider AI setup
Users can select different models and connect a personal API key through a clear, locally stored configuration.
04
Separated onboarding from daily use
The first run explains permissions and setup; later launches return the user directly to the active workflow.
05
Created the product site and docs
Landing, download, and documentation pages consistently explain value, compatibility, and first steps.
06
Prepared a cross-platform release
Shared states and a consistent visual system support macOS and Windows while keeping launch fast and predictable.
Tools
- React
- TypeScript
- Tailwind CSS