Rolebolt

AI-native recruiting and job-search workspace — hire faster with rubric-based scoring, pipeline automation, and a Gemini-powered copilot. Job seekers get match scores, application tracking, and career tools in one place.

Live product: rolebolt.tech
Judges testing kit: rolebolt.tech/recruit/judges

---

Overview

Rolebolt is a full-stack hiring platform built for small teams and growing companies. Recruiters create standard or structured form jobs, import resumes in bulk, review candidates with AI-assisted rubrics, automate pipeline actions, and collaborate with their team. Candidates browse public opportunities, see fit scores before applying, and manage their job search in a dedicated seeker workspace.

The product is AI-native: Google Gemini powers job description generation, resume scoring, form assessments, recruiting copilot conversations, daily briefings, and the Sitegen portfolio builder.

---

Features

For recruiters

Feature	Description
Standard jobs	Full ATS pipeline with stages, bulk resume import, and AI rubric scoring
Form jobs	Structured applications with async assessments and applicant timelines
AI Copilot	Ask questions about candidates, roles, and hiring data with grounded answers
Hiring Autopilot	Pipeline rules that automate repeatable actions (stage moves, emails)
What-If simulator	Model pipeline changes before committing
Private jobs	Publish roles visible only via direct link (hidden from Find Jobs)
Offers	Draft, version, and send candidate signing links
Collaboration	Team notes, activity feed, and shared candidate review
Analytics	Funnel metrics, source quality, and hiring pace
Billing	Free tier plus Razorpay subscriptions (Pro / Ultra)

For job seekers

Feature	Description
Find Jobs	Browse public opportunities with filters and job alerts
Match scores	See how your profile maps to a role before applying
Career workspace	Resumes, cover letters, application tracker, interview prep
Public profiles	Shareable seeker and creator profile pages

Additional products

• Sitegen — AI-generated portfolio websites from resume text (/website)
• Community reviews — Featured reviews, X posts, and video testimonials on the landing page

---

Tech stack

Layer	Technology
Frontend	Next.js 16, React 19, Tailwind CSS 4, TypeScript
Backend	Express 5, TypeScript, tsx runtime
Database	MongoDB (Mongoose)
AI	Google Gemini API, Mesh API (Gemini models), Nvidia NIM fallback
Auth	Custom JWT + email verification, Firebase (Google/Microsoft SSO)
Email	Resend
Payments	Razorpay
Deployment	Render (see render.yaml)

---

Project structure

.
├── backend/                 # Express API server
│   ├── src/
│   │   ├── ai/              # Gemini & Mesh clients
│   │   ├── automation/      # Pipeline rules & autopilot
│   │   ├── billing/         # Razorpay subscriptions & entitlements
│   │   ├── jobs/            # Cron jobs (briefings, offers)
│   │   ├── models/          # Mongoose schemas
│   │   ├── products/sitegen # Portfolio website builder
│   │   ├── recruit.ts       # Core recruiting routes
│   │   ├── recruitCopilot.ts
│   │   ├── recruitForms.ts
│   │   └── seeker.ts
│   └── .env.example
├── frontend/                # Next.js app
│   ├── src/
│   │   ├── app/             # App Router pages
│   │   ├── components/      # Shared UI components
│   │   ├── contexts/        # Auth & state
│   │   └── lib/             # API client, utilities
│   └── .env.example
├── Product_Evidence/        # Hackathon submission artifacts
├── render.yaml              # Render deployment blueprint
└── LICENSE

---

Getting started

Prerequisites

• Node.js 20.x (see .node-version in each package)
• MongoDB — local instance or Atlas connection string
• Google Gemini API key — for AI features
• Resend API key — for transactional email (optional in dev)

1. Clone and install

git clone <repo-url>
cd rolebolt

# Backend
cd backend
cp .env.example .env
# Edit .env with your values
npm install

# Frontend (new terminal)
cd ../frontend
cp .env.example .env.local
# Set BACKEND_URL=http://localhost:8080
npm install

2. Configure environment

Backend (backend/.env) — key variables:

Variable	Required	Purpose
MONGODB_URI	Yes	MongoDB connection string
SESSION_SECRET	Yes	JWT signing secret (64+ random bytes)
GEMINI_MESH_KEY	Yes*	Mesh API key for scoring, copilot, forms
GEMINI_PRIMARY_KEY	Yes*	Direct Gemini API for JD generation
FRONTEND_URL	Yes	Public frontend URL (email links)
CORS_ORIGIN	Prod	Allowed frontend origin(s)
RESEND_API_KEY	Prod	Email delivery
FIREBASE_SERVICE_ACCOUNT_JSON	SSO	Google/Microsoft sign-in
RAZORPAY_*	Billing	Payment processing

*At least one Gemini key is required for AI features to work.

Frontend (frontend/.env.local):

Variable	Purpose
BACKEND_URL	Backend API base URL
NEXT_PUBLIC_RECAPTCHA_SITE_KEY	Bot protection on applications
NEXT_PUBLIC_FIREBASE_*	Firebase client config for SSO

Generate a session secret:

node -e "console.log(require('crypto').randomBytes(64).toString('hex'))"

3. Run locally

# Terminal 1 — backend (port 8080)
cd backend
npm run dev

# Terminal 2 — frontend (port 5000)
cd frontend
npm run dev

Open http://localhost:5000.

4. Run tests

# Backend unit tests
cd backend && npm test

# Frontend unit tests
cd frontend && npm test

---

Deployment

The repo includes a Render Blueprint (render.yaml) that deploys two web services:

1. recruit-backend — Express API on port 8080
2. recruit-frontend — Next.js on port 5000

Set all sync: false environment variables in the Render dashboard. Point BACKEND_URL on the frontend to the deployed backend URL, and set CORS_ORIGIN / FRONTEND_URL on the backend to the frontend URL.

For Firebase SSO in production, paste the full service account JSON into FIREBASE_SERVICE_ACCOUNT_JSON — do not commit credential files.

---

AI architecture

User request
    │
    ├─► Gemini Primary API (GEMINI_PRIMARY_KEY)
    │       Job descriptions, structured generation
    │
    ├─► Mesh API / Gemini models (GEMINI_MESH_KEY)
    │       Resume scoring, copilot, form evaluation, briefings
    │
    └─► Nvidia NIM fallback (GEMINI_FALLBACK_KEY)
            Automatic fallback when primary models are unavailable

Health endpoints at /health and /ai-routing report AI service status.

---

Hackathon submission

Rolebolt is submitted to the Build with Gemini XPRIZE (Devpost).

• Category: Professional Services / Small Business Services
• Judges page: /recruit/judges — sample jobs, resumes, and walkthrough
• Judge's technical brief: JUDGES.md — Google Cloud stack, architecture diagrams, Gemini usage, roadmap
• Evidence folder: Product_Evidence/ — add GCP billing, Gemini usage screenshots, and revenue proof before final submission

---

Security notes

• Never commit .env files, firebase-service-account.json, or API keys
• RAKA98_ADMIN_PASSWORD secures the internal verification admin panel
• reCAPTCHA v3 protects public job applications
• Razorpay webhooks are verified with HMAC signatures

---

License

Proprietary — all rights reserved by Rolebolt.

This repository is public so Build with Gemini / XPRIZE judges can evaluate the submission. Public access is not an open-source license and does not grant the general public permission to use, copy, modify, fork, commercialize, or build another product from this code. See LICENSE.