Role Ready CV
AI-powered CV tailoring tool. Paste an existing CV and a job description, get a role-targeted, ATS-optimised version back in seconds. Built in Next.js, with the Anthropic Claude API as the tailoring engine.
Outcomes
Live AI product
Claude API integrated into a complete CV tailoring flow: paste a CV and a job description, get an ATS-optimised version back in seconds, with the original facts preserved.
Production-grade backend
Supabase Auth and Postgres with Row Level Security, Stripe Checkout with webhook reconciliation, and a credit-based ledger split across users and transactions tables.
Full document pipeline
Imports PDF (via unpdf) and .docx (via mammoth) and exports back to .docx and PDF, so the tool round-trips the formats job seekers actually have on disk.
No-subscription model
One-time credit packs via Stripe (free trial, then £5 for 5 credits or £9 for 15). No recurring billing, no dark patterns, no surprise charges.
The story
Job seekers know they should tailor a CV for every role. The friction stops them. Each application takes 30 to 60 minutes of rewording, context-switching between their own experience and each role's language, and the output is still likely to be filtered out by an ATS that doesn't see the right keywords. Most people give up and send a generic version.
Role Ready CV closes the loop in three steps: paste an existing CV, paste a target job description, get back an ATS-optimised version in seconds. The AI restructures and rephrases what's already there to match what the role actually asks for. It does not fabricate experience.
AI behaviour
The tailoring engine is built on Anthropic's Claude API with a tightly constrained prompt. Claude must preserve the factual accuracy of the original CV (no invented roles, dates, or skills), optimise wording for relevance to the target role, emphasise genuinely matching skills and experience, and surface keywords the ATS is likely to look for. If the gap between the source CV and the target role is too wide, the model says so rather than papering over it.
Stack
Next.js 16 on the App Router, TypeScript end to end, Tailwind for styling.
Supabase Auth handles email-based authentication; user records, credit
balances, and transaction history live in a Postgres schema with Row Level
Security so users can only ever see their own data. A Supabase database
trigger auto-creates the users row on signup and seeds a single free
credit so the first tailored CV is on the house.
Stripe Checkout drives one-time credit purchases, and a webhook handler
reconciles successful sessions into the transactions table and credits
the user atomically. No subscriptions, no recurring billing, no dunning
flow to maintain.
Document pipeline
The CV input doesn't have to be plain text. Role Ready imports PDF (via
unpdf) and .docx (via mammoth), extracts the text content, runs the
tailoring pass, and exports the result back to .docx (via docx) and PDF
(via html2pdf.js), all client-side. Job seekers can drop in the file
they already have and download the tailored version in the same format.
Status
Live at rolereadycv.co.uk under the Designspin brand. Free trial credit on signup, then one-time top-ups via Stripe.