Tailor every resume.
Apply on your terms.
Paste a job URL. Rendure scrapes the posting, tailors your resume, runs an automated QA pass, and stores every run in Postgres. It never submits anything — you review the version and the QA notes, then apply yourself.
A tailoring pipeline you actually own.
Rendure turns a job posting into a tailored resume — and a paper trail you can audit.
Point it at a job URL and a chain of small, single-purpose agents takes over: one scrapes the posting, one rewrites your resume against it, one grades the result, and a final one hands you a version to review. Each agent does exactly one job, writes what it learned to the database, and steps aside for the next.
There's no shared cloud and no account to create. It runs on your machine, talks to language models through your own OpenRouter key, and records every job, resume version, QA score, and pipeline event in a Postgres database you control — nothing tucked away in a vendor's backend.
Crucially, it stops at the resume. Rendure never clicks "apply." It does the tedious tailoring and quality checks, then leaves the judgment — and the submit button — to you.
- Runtime
- Python 3.12
- Storage
- PostgreSQL
- Models
- via OpenRouter
- Interface
- Web + SSE
- Export
- RenderCV PDF
- Deploy
- Docker Compose
- License
- MIT
How it works
A four-agent loop runs the whole thing — streamed to your screen over server-sent events. No black box: every step is logged, scored, and stored.
Every draft is scored, not guessed.
The Quality Analyst grades each tailored resume against the job description across four dimensions. A draft only passes when the weighted composite clears the threshold and the structure check holds — otherwise it loops back for another tailoring pass.
Fraction of the JD's keywords and required skills that actually appear in the resume.
How many of the role's responsibilities are backed by real bullet points.
Whether the tone and scope of the resume align with the role's level.
All required sections present, no placeholders. A break here forces the whole score to zero.
Built to be boring & honest.
Rendure never applies for you.
This is the boundary, and it's deliberate. Rendure prepares tailored resume versions and QA notes — nothing leaves your machine, and nothing gets submitted to an employer. You read the diff, you check the gaps, you decide where and when to apply. The robot does the busywork; the judgment stays yours.
Everything runs on your box.
Encrypted key storage
Your OpenRouter API key is encrypted at rest. It powers the agents and never leaves your host.
Postgres audit trail
Every run, version, score, and gap is written to Postgres. Full history, queryable, yours forever.
Live SSE progress
Watch each agent work in real time over server-sent events — the same feed that streams to the dashboard.
RenderCV PDF export
Approved versions render to clean, ATS-safe PDFs via RenderCV. Typeset like a document, not a webpage.
Telegram notifications
Optional pings when a run finishes or a QA pass fails. Send /start to your bot and paste the chat ID.
Multi-agent pipeline
Scout, Tailor, Quality Analyst, and Confirmation — a Python pipeline you can read, fork, and extend.
The short version.
Does Rendure submit applications for me?
Where does my data live?
Which language models does it use?
Will it fabricate experience to match the job?
How do I run it?
Clone it. Run it. Own it.
No accounts, no SaaS dashboard, no waitlist. Pull the repo, set your key, and bring the whole pipeline up with one command.