Resume Signal

Resume Signal is a MERN resume analyzer and ATS checker. It accepts PDF or DOCX resumes, parses recruiter-facing sections, scores them against a job description, stores report history, and enriches the report with OpenAI, Gemini, or Groq when a provider key is configured.

What ships

• React/Vite SaaS UI with Tailwind, Framer Motion, Redux Toolkit, React Router, Dropzone upload, Recharts dashboards, responsive dark mode, report history, and admin views.
• Express/Mongo API with JWT cookies and bearer tokens, Google identity login, forgot/reset password flow, Multer memory uploads, PDF/DOCX text extraction, Mongoose models, validation, Helmet, CORS, rate limits, sanitization, Winston logging, and optional Redis AI-result caching.
• ATS pipeline for keyword match, target skill gaps, formatting, completeness, readability, ATS compatibility, strength/weak-area summaries, and deterministic fallback guidance.
• AI provider adapters for openai, gemini, groq, and a mock fallback. AI covers review, summary, bullet enhancement, ATS suggestions, projects, career suggestions, skill gap analysis, interview questions, and rewrite suggestions.
• Docker Compose, API/web Dockerfiles, Render blueprint, Vercel SPA rewrite config, GitHub Actions CI, environment examples, seed data, and a scoring test.

Architecture

.
|-- apps/
|   |-- api/
|   |   |-- src/config       # env, Mongo, Redis, logging
|   |   |-- src/models       # User, Resume, Analysis, JD, Feedback
|   |   |-- src/routes       # auth, analyses, admin, feedback
|   |   |-- src/services     # parsing, ATS scoring, AI, email
|   |   `-- scripts/seed.js
|   `-- web/
|       `-- src             # app store/router, pages, components, API client
|-- .github/workflows/ci.yml
|-- docker-compose.yml
`-- render.yaml

Local setup

1. Use Node 20+, run corepack enable, and run yarn install from the repo root.
2. Copy apps/api/.env.example to apps/api/.env.
3. Copy apps/web/.env.example to apps/web/.env.
4. Start MongoDB and Redis locally, or run docker compose up mongo redis.
5. Run yarn dev.

The web app defaults to http://localhost:5173; the API defaults to http://localhost:5000/api/v1.

Seed demo data with yarn seed. The seed user is demo@resumesignal.dev with password DemoPassword123!; change or remove that account outside local demos.

Environment

API essentials:

Variable	Purpose
MONGODB_URI	MongoDB Atlas or local Mongo connection
JWT_SECRET	At least 32 random characters
CLIENT_URL	Allowed frontend origin
AI_PROVIDER	openai, gemini, groq, or mock
REDIS_URL	Optional Redis cache
GOOGLE_CLIENT_ID	Google identity token audience
SMTP_*	Optional password reset mail transport

Frontend essentials:

Variable	Purpose
VITE_API_URL	Versioned backend base URL
VITE_GOOGLE_CLIENT_ID	Enables the Google sign-in button

Only one AI provider key is needed. Without one, scoring and stored reports still work with deterministic guidance.

API surface

Method	Route	Purpose
POST	/api/v1/auth/signup	Email/password signup
POST	/api/v1/auth/login	JWT session
POST	/api/v1/auth/google	Verify Google identity credential
POST	/api/v1/auth/forgot-password	Email reset request
POST	/api/v1/auth/reset-password	Set password with reset token
GET	/api/v1/auth/me	Current user
POST	/api/v1/analyses	Multipart resume plus job description
GET	/api/v1/analyses	Report history
GET	/api/v1/analyses/dashboard	Chart payloads
GET	/api/v1/analyses/:id	One owned report
POST	/api/v1/feedback	Store report feedback
GET	/api/v1/admin/overview	Admin analytics and feedback

The analysis upload field is resume; accepted files are PDF and DOCX up to 5 MB.

Deployment

• Web: import apps/web into Vercel or use the root workspace build command yarn workspace @resume-signal/web build. Set VITE_API_URL before build.
• API: render.yaml builds apps/api/Dockerfile; Railway can use the same Dockerfile or yarn workspace @resume-signal/api start.
• Data: use MongoDB Atlas. Redis is optional but recommended for repeated AI analyses.
• Containers: copy env files and run docker compose up --build.
• CI: GitHub Actions installs, tests ATS scoring, and builds both workspaces.

Production notes

Use a strong JWT_SECRET, HTTPS origins, restricted Google OAuth origins, SMTP credentials from a secret manager, provider API keys with usage budgets, MongoDB indexes/backups, and external observability for logs/alerts. The deterministic ATS score is guidance, not a promise about a specific employer's ATS.