Job Search Copilot

A local-first job search management app built with Next.js, React, and Tailwind, with AI copilot architecture implemented and disabled by default.

The product is designed to support a practical end-to-end workflow:

Jobs -> Recruiters -> Interviews -> Follow-ups -> Activity Timeline -> Insights -> Backup/Restore

Live Demo

https://ai-job-search-app.vercel.app

Repository

https://github.com/AliaksandrHv/ai-job-search-app

Changelog

See CHANGELOG.md for release history and semantic versioning notes.

Demo

• Live app: https://ai-job-search-app.vercel.app
• Walkthrough video (YouTube): https://youtu.be/myvHFwZlb8s

Suggested walkthrough order:

1. Open dashboard
2. Add a job
3. Edit the job
4. Delete a job
5. Show filtering and search
6. Show recruiter tracking
7. Refresh to confirm persistence

Problem

Managing job applications across multiple companies can become scattered and hard to track. This project was built to provide one operational dashboard for applications, recruiters, interviews, follow-ups, and recent activity.

Screenshots

Dashboard

Applications

Interviews

MVP Features

Jobs

• Add, edit, delete jobs
• Required validation for company and title
• Status tracking (Saved, Applied, Interview, Rejected)
• Optional job link support
• Follow-up fields and actions

Recruiters

• Add, edit, delete recruiter contacts
• Search and sort recruiters
• Track contact dates and next follow-up
• Link/unlink recruiter to a job

Interviews

• Add, edit, delete interviews
• Link interviews to jobs and optionally recruiters
• Status tracking (Scheduled, Completed, Cancelled)
• Filters, search, and sort modes
• "Needs outcome update" detection for past scheduled interviews

Follow-ups

• Job-level follow-up status and dates
• Quick actions:
  • set/reschedule follow-up
  • mark contacted today
  • mark follow-up done
  • clear follow-up
• Dashboard cards for due/overdue visibility

Activity Timeline

• Auto-logs key product actions
• Dashboard "Recent Activity" widget
• Job Details and Recruiter Details timelines
• Persists in localStorage with cap to avoid unbounded growth

Backup / Restore

• Export full app data to JSON
• Import JSON backup with validation and overwrite confirmation
• Immediate UI refresh after successful restore

AI Copilot

• Feature-flagged AI service layer and secure server routes
• Job description summarizer
• Smart job notes generator
• Follow-up message generator
• Disabled by default in the public build to avoid API cost

Tech Stack

• Next.js 16 (App Router)
• React 19
• TypeScript
• Tailwind CSS v4
• localStorage persistence
• Next.js server routes for optional AI requests

Tech Decisions

• Next.js: fast product iteration with a production-ready framework
• Tailwind CSS: rapid UI development with consistent design primitives
• localStorage: simple persistence for MVP speed and offline-friendly behavior
• Playwright: end-to-end coverage for core user flows

Architecture Overview

Current implementation is intentionally single-page and local-first:

• Main UI and state orchestration: app/page.tsx
• Global styles/theme tokens: app/globals.css
• App shell metadata: app/layout.tsx
• AI routes and provider layer: app/api/ai/*, lib/ai/*

System view:

User
  -> Next.js Frontend
    -> React State
      -> localStorage
    -> AI Feature Flag
      -> Next.js AI Routes
        -> OpenAI API (optional, disabled by default)

Data lifecycle:

1. Hydrate state from localStorage on app load
2. Normalize records to keep backward compatibility
3. Persist each entity set through dedicated useEffect syncs
4. Render views (Dashboard, Applications, Recruiters, Interviews)
5. Optionally call AI routes for summaries, notes, and follow-ups when enabled
6. Open detail/edit panels for focused record operations

Data Model (MVP)

Job

• id
• company
• title
• status
• note
• jobDescription?
• jobLink?
• linkedRecruiterId?
• nextFollowUpDate?
• lastContactDate?
• followUpStatus?
• aiSummary?
• aiNotes?
• aiFollowUpSuggestion?
• aiUpdatedAt?
• createdAt?

Recruiter

• id
• name
• company
• role
• email
• profileLink?
• notes
• lastContactDate?
• nextFollowUpDate?
• createdAt?

Interview

• id
• jobId
• recruiterId?
• interviewType
• status
• scheduledAt
• notes
• createdAt?
• updatedAt?

ActivityEvent

• id
• type
• timestamp
• message
• jobId?
• recruiterId?
• interviewId?

Persistence Keys

• ai_job_search_jobs_v1
• ai_job_search_recruiters_v1
• ai_job_search_interviews_v1
• ai_job_search_activity_v1

AI Configuration

• Feature flag: NEXT_PUBLIC_FEATURE_AI_COPILOT
• Provider secret: OPENAI_API_KEY
• Default behavior: AI stays hidden and inactive unless explicitly enabled
• Security: AI requests run through app/api/ai/*; the API key is never exposed client-side

Run Locally

npm install
npm run dev

Open http://localhost:3000.

Available Scripts

npm run dev
npm run lint
npm run build
npm run start

Testing

End-to-end tests are implemented using Playwright.

Run tests locally:

npm run test:e2e

Covered flows:

• Add job
• Edit job
• Delete job
• Search jobs
• Filter by status

CI automation:

• GitHub Actions workflow: .github/workflows/e2e-playwright.yml
• Runs on every push and pull request to main
• Uploads Playwright artifacts (playwright-report/ and test-results/)
• Includes recorded browser sessions as test video artifacts

Roadmap

v1.1

• backend persistence and authentication
• shared data across devices

v1.2

• resume vs job match scoring
• interview question generation support

v1.3

• recruiter follow-up reminders
• deeper analytics and trend tracking

Project Structure (Current)

app/
  api/ai/
  globals.css
  layout.tsx
  page.tsx
components/
  ai/
config/
  features.ts
lib/
  ai/

Portfolio Notes

This project demonstrates:

• Product-driven feature sequencing from MVP to usability polish
• Strong local data modeling and backward-compatible persistence
• Connected workflow across entities (jobs, recruiters, interviews, follow-ups)
• Operational UX (dashboard insights, details panels, quick actions)
• Feature-flagged AI architecture with server-side provider isolation
• Practical data portability via JSON backup/restore

MVP Status

MVP is complete and functional as a standalone local-first product.

Potential next iterations:

• Modular code split (components/, hooks/, lib/, types/)
• richer keyboard shortcuts
• priority/pinned jobs
• enhanced analytics widgets