{ Private AI Interview Practice }

PRACTICE INTERVIEWS WITH PRIVATE AI

Private, realistic voice interview practice with honest scoring that stays on your device.

Start practice
★ GitHub (Samyak9109/Abhyas) ↗Created by Samyak Jain
{ Why choose us }

BUILT FOR REAL PROGRESS

01 // SPEED

Fast start

Jump straight into realistic practice in under two minutes. No account creation, no email verification, and no payment walls.

02 // PSYCHOLOGICAL SAFETY

Safe space

Practice high-stakes behavioral and system design questions in private comfort without judgment, awkward pauses, or fear of failure.

03 // ACTIONABLE FEEDBACK

Track progress

Honest, grounded rubric scoring across technical depth, communication clarity, and delivery pace with concrete rewrites.

04 // VERIFIABLE SOVEREIGNTY

Private and local

Your resume, answers, and audio never touch our servers. Everything stays in your browser's IndexedDB or runs completely offline.

{ The experience }

PRACTICE THAT FEELS LIKE THE REAL THING

Works offline once a local model is downloaded
Senior Staff Interviewer // Speaking
"Can you describe a challenging technical trade-off you had to make in your recent system architecture?"
[PUSH TO TALK]Live microphone speech recognition active...
  • Voice push-to-talk & real-time subtitles: Natural conversational rhythm with instant editable transcription.
  • Full voice and text equality: Everything you can do by speaking, you can do by typing.
  • Dynamic follow-up questions: Probes depth and detects vague responses just like experienced human interviewers.
  • Top three actionable fixes first: Clear, prioritized takeaways so you know exactly where to improve next.
Need an API key or prefer running local models? Read our setup guide →
{ The facts }

VERIFIABLE GUARANTEES

0ACCOUNTS NEEDED

No signups, no passwords, no email collection, and no tracking cookies.

0RESUMES SENT TO OUR SERVERS

All PDF and Word documents are parsed strictly inside your browser via WebAssembly.

MITOPEN SOURCE LICENSE

Inspect every line of code, verify network egress, or self-host your own instance.

{ How it works }

FOUR STEPS TO PREPARATION

01

Connect your AI

Enter your Google Gemini or OpenAI API key, or connect to a local runner. Keys stay strictly in your browser.

02

Add your profile

Upload or paste your resume and target job role. Skills and experience are parsed locally in milliseconds.

03

Choose your session

Select question difficulty, interview focus, interviewer style, and choose voice push-to-talk or text mode.

04

Practice & improve

Answer realistic questions, receive follow-ups, and review rubric scores with concrete, grounded rewrites.

{ Common questions }

FREQUENTLY ASKED QUESTIONS

Why choose Abhyas over other interview prep platforms?
Most interview platforms charge costly subscriptions, store your personal resumes and speech recordings on their central servers, and provide generic “good job” praise. Abhyas is completely free, 100% open-source, and architected around client-side privacy: your data and audio never touch an intermediary server. Moreover, Abhyas provides rigorous, rubric-calibrated scoring with grounded rewrites that never invent facts absent from your profile.
How is Abhyas better than commercial AI interview competitors?
Commercial tools lock you into a single proprietary AI model, force account registration, and gate realistic mock sessions behind paywalls. Abhyas gives you sovereignty with Bring-Your-Own-Key (BYOK) support across Gemini, OpenAI, Claude, Hugging Face, and local offline engines (Ollama, WebLLM). You get honest level-calibrated difficulty, skill streak analytics without guilt resets, evidenced company packs, and calendar spaced repetition — all running locally in your browser.
Is Abhyas free to use?
Yes, Abhyas is completely free and open-source under the MIT license. You bring your own API key (BYOK) from providers like Google Gemini (which offers a generous free tier) or OpenAI, or run completely free offline models.
Where does my resume and interview data go?
Nowhere except directly to your chosen AI provider. Abhyas operates without an intermediary backend server. Your resume text, answers, and evaluations reside strictly inside your browser's local IndexedDB.
Can I switch between voice and text during an interview?
Yes. Voice and text are equal first-class citizens. You can speak your answer using push-to-talk or on-device hands-free voice activity detection (VAD), edit the transcribed text before submitting, or switch entirely to typing at any point.
Does Abhyas work completely offline?
Yes, once the static web assets are loaded, you can run drills with question banks offline. Furthermore, connecting to WebGPU in-browser models or a local Ollama server on your machine allows complete interview sessions with zero internet connection.
How are my API keys protected?
API keys are stored exclusively in your browser's local IndexedDB with optional WebCrypto AES-GCM passphrase encryption. They are never sent to any Abhyas server, never placed in URLs, and never logged.

Practice today. Be ready tomorrow

Free, open-source, and completely private interview practice.

Start practice