Built to solve recruiter sourcing bottlenecks without expensive LinkedIn Recruiter licenses or manual profile screening.
Recruiters spend 60–70% of their time manually screening profiles, categorizing skills, and crafting personalized outreach messages. LinkedIn Recruiter licenses cost $10,000+/year, and most ATS tools don't automate the sourcing workflow itself.
GigWeave is an AI sourcing tool that classifies freelancer profiles by skill category with evidence-based confidence scoring, extracts experience signals, and generates targeted outreach messages — all client-side with zero API costs per scan. Paste a profile or upload a screenshot, get a classified candidate and a draft message in under 3 seconds.
📹 Demo: [Insert 30-second screen recording showing: paste profile → instant classification → generated outreach message]
🔗 Deployed on Vercel
- Classification speed: Sub-500ms profile analysis — deterministic rule engine requires no LLM API calls
- Zero operating cost: Every classification runs client-side in the browser — no API fees, no GPU, no servers
- Evidence-based confidence: Every skill match includes the source text quote — auditors can verify every classification
- 9 skill categories: Web, Mobile, AI/ML, Data, DevOps, Design, Writing, Marketing — with intelligent tie-breaking rules
- Dual output format: Generates both email drafts (full context) and LinkedIn messages (300-character optimized)
- Vision pipeline: Drag-and-drop screenshot/PDF upload simulates OCR-based profile extraction for mobile-first sourcing
- X-Ray search: Generates Google dork queries to find LinkedIn profiles anonymously — bypasses LinkedIn search limits
| Layer | Technology |
|---|---|
| Frontend | Next.js 16 (App Router, Turbopack), TypeScript |
| Animation | framer-motion |
| UI | Tailwind CSS v4, lucide-react icons |
| Classification Engine | Deterministic rule-based (regex + category precedence) |
| Deployment | Vercel (serverless) |
Profile Text (paste or screenshot)
│
▼
Category Detection ──→ Regex matches against 9 skill taxonomies
│
▼
Tie-Breaker Engine ──→ Priority rules (Mobile > Web, AI/ML > General, etc.)
│
▼
Evidence Extraction ──→ Quotes source text for every classification
│
▼
Confidence Scoring ──→ HIGH / MEDIUM / LOW based on signal strength
│
▼
Message Generation ──→ Email or LinkedIn-optimized draft
pnpm install
pnpm devOpens at http://localhost:3000.
MIT — see LICENSE