First Cut — One-Pager

The Problem

01

Recruiters were spending 2–3 hours per role reading resumes before any screening conversation happened. Most were clear rejects within the first 10 seconds.

02

Inbound from job boards and career site mixed with referrals and outbound responses. No consistent scoring. Different recruiters applied different bars.

03

Rejected candidates got no response. Good candidates with timing or location issues got lost instead of tagged for future roles.

04

Shortlisting decisions lived in someone's head. No audit trail, no documented reason, nothing to learn from when a hire worked out or didn't.

How It Works

01

Set the JD

Paste the job description, set the role, domain, years of experience required. Add must-haves and auto red-flag conditions.

›

02

Upload Resumes

Drag up to 100 PDFs at once. Text-based and scanned PDFs both work — Claude vision handles OCR fallback automatically.

›

03

Claude Scores

Each resume gets an Experience score and Fit score, both 0–10. Total out of 20. Plus red flags, a 3-sentence hiring-manager brief, and a verdict.

›

04

Push to Pipeline

Results push to Hiring Intelligence via the central upsert endpoint. Strong Yes → Shortlisted. Maybe → On Hold. Reject → Rejected. Deduped automatically.

The Scoring Framework

Experience (0–10)

Years in relevant domain

Seniority progression

Industry and company type

Exact title match vs adjacent

Fit (0–10)

Specific tools and methods named

Types of projects delivered

Stakeholder level and team size

Measurable outcomes vs vague claims

Verdicts

Strong Yes

Total ≥ 15, no critical constraint missing

Reject

Total ≤ 7 or must-have absent

Maybe

Genuine uncertainty — human call needed

What Gets Extracted Per Resume

👤

Name · Email · Phone · Location

🎓

Education · College · No bootcamps

🔗

LinkedIn URL · Deep search in email + resume

📋

Summary · Red flags · Scores · Verdict

"Candidate has 3 years listed but graduated 2 years ago — timeline doesn't add up. VD background, not primary UX research. Portfolio shows mostly UI, nothing on research methods."

That's the summary Claude writes. Specific to the candidate, specific to the JD. Not a template.

What Makes This Different

01

Scores are calibrated, not vibes

Experience (0–10) and Fit (0–10) are scored against explicit rubrics in the prompt. 9–10 means direct domain match at the right seniority. 5–6 means adjacent experience. Every recruiter gets the same bar, every time.

02

Scanned PDFs work too

Most screening tools fail on image-based PDFs. First Cut sends scanned resumes to Claude vision as base64 documents and reads them natively. An OCR badge flags the card so the recruiter knows the confidence level.

03

Feeds directly into Hiring Intelligence

Every scored resume pushes to the hiring pipeline via the central upsert endpoint — not a separate sheet, not a CSV export. Scores, verdicts, and rejection notes land in the Daily Tracker instantly. The Talent Pool engine uses this data to match soft-rejects to future roles.

04

Auto-reply with signature images

Upload your signature image once, it persists across all outgoing emails. Select rejected candidates, customize the template, and send via Gmail API in one click. Candidates get a professional response with your branding. No copy-pasting, no manual signatures.

05

Enhanced LinkedIn extraction

Searches thoroughly across email body, signature, resume header, and contact sections. Finds LinkedIn URLs in multiple formats. No more missed connections — comprehensive candidate data captured automatically from day one.

Smart Response Templates

Strong Yes

Personalized next-steps email. Schedule interview, introduce the role, set expectations.

Maybe

Still evaluating. Request more info or portfolio. Timeline uncertain — keep candidate warm.

Reject

Professional rejection. Point to careers page. Encourage future applications.

No Opportunities

Generic rejection. No relevant openings for candidate skills. Stay in touch for future roles.

No Internships

Specific response for internship inquiries when none available. Encourage reapplication for full-time roles.

All templates are fully editable, saveable as custom versions, and sent with your signature image automatically.

Connected to Hiring Intelligence

First Cut feeds the pipeline

Every candidate scored in First Cut becomes a row in the Daily Tracker. Stage maps directly: Strong Yes → Shortlisted, Maybe → On Hold, Reject → Rejected. The AI Score (0–20) and Verdict columns in the sheet are written automatically. When the Talent Pool engine runs, it reads these scores to identify silver-medal candidates worth re-engaging for future roles. Email replies are sent with professional signature images.

Stack

Netlify Cloudflare Worker Claude Sonnet · Vision + Text Google Sheets API Gmail API · bulk reply Google OAuth · @netbramha.com only PDF.js · OCR fallback Single HTML · no build step