SonicPrep

Practice the conversation. Understand the feedback. Walk into the real interview prepared.

Product preview · Run locally · Quality checks

What it does

SonicPrep is a responsive interview practice app. Candidates prepare a role-specific interview, speak with an AI interviewer, and turn feedback into a practical improvement plan.

Step	Experience
Prepare	Create an interview for a role, experience level, and technology stack. Optionally upload a PDF resume for questions about your experience. Company interviews display the matching company cover where available, including Amazon.
Practice	Choose a male or female candidate avatar before the voice session. Follow the live transcript while speaking with the AI interviewer.
Review	See an overall score, five assessment categories, strengths, specific improvements, and a final assessment. Retake interviews and revisit your history.
Take away	Download a two-page PDF with the assessment, prioritized next steps, an improvement mind map, and role-specific technical focus.

Email/password authentication, password reset, community interviews, and light and dark themes complete the experience. The interview view keeps SonicPrep branding in the navigation while the selected company cover identifies the session.

Optional resume questions

On Create an interview, enter your target role, experience level, skills, interview type, and question count. You can add a PDF resume (up to 4 MB). SonicPrep reads the PDF once and uses its projects, skills, and achievements when generating role-specific questions. The first question names a resume detail. The interview is saved before the app opens voice practice, so a voice connection issue cannot prevent creation. The PDF is not stored; the temporary extracted text is deleted after questions are saved. Retries reuse the extracted resume within the same setup and recover an already saved interview without creating a duplicate. Incomplete extraction is rejected with a clear message. Leave the resume field empty for a standard interview. This uses the existing GOOGLE_GENERATIVE_AI_API_KEY; no extra service or key is needed.

Product preview

The screenshots use fictional sample data, including an Amazon SEO Specialist interview. Select an image to see its full capture. The creation form saves the interview before voice practice begins. These previews do not show live Firebase, Gemini, or Vapi calls.

Dashboard

Desktop	Mobile

Create an interview

Choose a role, experience level, skills, and question count. Add an optional PDF resume for questions drawn from your experience.

Desktop	Mobile

Continue with voice practice and a live transcript once the interview is saved.

Feedback and report

Desktop	Mobile

Assessment	Improvement plan

Account access: desktop sign-in · mobile sign-in. See screenshot notes for capture details.

Technology

Layer	Tools
App and interface	Next.js 16 App Router, React 19, TypeScript, Tailwind CSS 4, Radix UI
Forms	React Hook Form, Zod
Accounts and data	Firebase Authentication, Firebase Admin, Firestore
Voice	Vapi Web SDK
Questions and feedback	Google Gemini through the AI SDK
Resume PDF processing	Google Gemini through the same AI SDK and key

Run locally

Use Node.js 22.13+ on the 22.x line or 24+, and npm. Firebase, Vapi, and Google AI credentials are needed for the live interview flow.

1. Install dependencies with npm ci.

2. Copy .env.example to .env.local and add your credentials:

   Variable	Used for
   NEXT_PUBLIC_FIREBASE_*	Firebase web app
   FIREBASE_PROJECT_ID, FIREBASE_CLIENT_EMAIL, FIREBASE_PRIVATE_KEY	Firebase Admin
   GOOGLE_GENERATIVE_AI_API_KEY	Gemini
   NEXT_PUBLIC_VAPI_WEB_TOKEN	Voice interview practice
   NEXT_PUBLIC_VAPI_WORKFLOW_ID, VAPI_WEBHOOK_SECRET	Legacy voice setup webhook, if still used

3. Enable Email/Password in Firebase Authentication, authorize your domain, and create a Firestore database.

4. Configure the Vapi web token for voice practice. New interviews are created and saved in the app before the voice call starts; the legacy Vapi generation workflow is no longer required for creation.

5. Run npm run dev and open localhost:3000. Allow microphone access for voice practice.

Keep .env.local private. Production builds do not require live credentials, but connected interview features do.

Quality checks

Command	Check
npm run check	ESLint, TypeScript, and regression tests
npm run format:check	Formatting
npm run build	Production compilation
npm start	Serve the production build

Author and license

Created by Salony Ranjan · GitHub · LinkedIn

Licensed under the MIT License.