HowlerHire

AI-powered job search and resume optimization application. Automatically finds matching jobs, generates tailored resumes and cover letters, and tracks your applications.

Tech Stack

Layer	Technology
Frontend	SvelteKit + shadcn-svelte
Database	Supabase (PostgreSQL + pgvector + Auth + RLS)
LLM Gateway	LiteLLM (self-hosted, multi-provider)
Workflows	Inngest (durable execution)
Job Sources	Adzuna, The Muse, Greenhouse APIs
Deployment	Cloudflare Pages + Workers

Features

• Daily Job Discovery: Automatically finds 20 matching jobs per day
• Smart Matching: Keyword-based job matching with skill analysis (embedding support in schema)
• Tailored Documents: AI-generated resume and cover letter for each job
• Application Tracking: Mark jobs as applied, saved, or not relevant
• Learning System: Improves recommendations based on your feedback
• Multi-provider LLM: Switch between Claude, GPT, and Gemini as needed

Quick Start

Prerequisites

• Node.js 20+
• Supabase account (free tier works)
• LLM API keys (at least one of: Anthropic, OpenAI, Google)
• Inngest account (free tier: 50k executions/month)

1. Clone and Install

git clone <repo-url>
cd howlerhire
npm install

2. Set Up Supabase

1. Create a new Supabase project at https://supabase.com
2. Go to SQL Editor and run the migration:

   cat supabase/migrations/0001_initial_schema.sql

3. Enable the vector extension in Extensions settings
4. Copy your project URL and keys

3. Configure Environment

cp .env.example .env

Edit .env with your credentials:

# Supabase
PUBLIC_SUPABASE_URL=https://your-project.supabase.co
PUBLIC_SUPABASE_ANON_KEY=your-anon-key
SUPABASE_SERVICE_ROLE_KEY=your-service-role-key
SUPABASE_DB_URL=postgresql://postgres.[project-ref]:[password]@aws-0-[region].pooler.supabase.com:6543/postgres

# LLM API Keys (at least one required)
ANTHROPIC_API_KEY=sk-ant-xxx
OPENAI_API_KEY=sk-xxx
GOOGLE_API_KEY=xxx

# Job APIs (optional, for broader search)
ADZUNA_APP_ID=xxx
ADZUNA_APP_KEY=xxx

# Inngest
INNGEST_EVENT_KEY=xxx
INNGEST_SIGNING_KEY=xxx

4. Start Development

# Start the dev server
npm run dev

# In another terminal, start Inngest dev server
npm run inngest:dev

Open http://localhost:5173

5. (Optional) Run LiteLLM Proxy

For multi-provider LLM access with cost tracking:

# Start all services
docker-compose up -d

# Or just LiteLLM
docker run -p 4000:4000 \
  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
  -e OPENAI_API_KEY=$OPENAI_API_KEY \
  ghcr.io/berriai/litellm:main-latest

Then set LITELLM_PROXY_URL=http://localhost:4000 in your .env.

Project Structure

howlerhire/
├── src/
│   ├── lib/
│   │   ├── components/ui/     # shadcn-svelte components
│   │   ├── server/
│   │   │   ├── database/      # Drizzle schema & types
│   │   │   ├── inngest/       # Durable workflows
│   │   │   ├── jobs/          # Job API integrations
│   │   │   └── llm/           # LLM client & generators
│   │   └── utils.ts
│   ├── routes/
│   │   ├── auth/              # Login, signup, OAuth
│   │   ├── dashboard/         # Main app UI
│   │   │   ├── jobs/          # Job listing & details
│   │   │   ├── profile/       # User profile
│   │   │   └── resumes/       # Resume management
│   │   └── api/               # API endpoints
│   └── app.css                # Tailwind styles
├── supabase/migrations/       # Database migrations
├── docker-compose.yml         # Local services
└── wrangler.toml              # Cloudflare config

Workflows

Daily Job Discovery (Cron: 6 AM UTC)

1. Fetch user profile and preferences
2. Search all job sources (Adzuna, Muse, Greenhouse)
3. Filter duplicates and rank by relevance
4. Save top 20 matches to database
5. Queue resume generation for each job

Resume Generation

1. Load job details and user profile
2. Generate tailored resume via LLM
3. Generate cover letter via LLM
4. Save to database and notify user

Deployment

Cloudflare Pages

# Install Wrangler CLI
npm install -g wrangler

# Login to Cloudflare
wrangler login

# Deploy
npm run build
wrangler pages deploy .svelte-kit/cloudflare

# Set secrets
wrangler pages secret put SUPABASE_SERVICE_ROLE_KEY
wrangler pages secret put ANTHROPIC_API_KEY
# ... etc

Environment Variables

Set these in Cloudflare dashboard > Pages > Settings > Environment variables:

• PUBLIC_SUPABASE_URL
• PUBLIC_SUPABASE_ANON_KEY
• SUPABASE_SERVICE_ROLE_KEY
• ANTHROPIC_API_KEY (or other LLM keys)
• INNGEST_EVENT_KEY
• INNGEST_SIGNING_KEY

Cost Estimates

Free Tier Usage

• Supabase: Free (500MB DB, 50K MAU)
• Inngest: Free (50K-100K executions/month)
• Cloudflare: Free (100K requests/day)
• Job APIs: Free (Adzuna, Muse, Greenhouse)

LLM Costs (estimated per month)

• 20 jobs/day × 30 days = 600 jobs
• Resume + Cover letter per job ≈ $0.05-0.15
• Total: ~$30-90/month in LLM costs

Adding More Job Sources

To add a new job API, create a new file in src/lib/server/jobs/:

// src/lib/server/jobs/newapi.ts
import type { JobResult, JobSearchParams } from './types';

export async function searchNewApiJobs(params: JobSearchParams): Promise<JobResult[]> {
  // Implement search logic
  return [];
}

Then add it to the job discovery workflow in src/lib/server/inngest/functions/job-discovery.ts.

TODO

• Add Lever API integration (implemented in src/lib/server/jobs/lever.ts)
• Add Apify/Scrapingdog for broader coverage
• Implement embedding-based job matching (schema ready with pgvector)
• Add resume upload and parsing (implemented via LLM extraction)
• Add email notifications
• Add analytics dashboard

License

MIT