Rungleaf is a resume and cover-letter builder that you host yourself, built by DEV3 Technologies on Next.js 16, React 19, TypeScript, Tailwind 4 and MongoDB, with a separate Node worker that runs the AI queue. Individuals get a personal workspace. Career coaches, agencies and college career offices get a team workspace with client seats. People build a resume from structured sections, pick one of twelve templates, edit beside a live page preview, tailor it to a pasted job description, write a cover letter, and send a PDF or a private link. You, the operator, sell plans and credit packs and run the platform from an operator panel. The demo shows fictional people and employers only.
How the AI works, plainly: Rungleaf ships Gemini AI admin tools: 17 assist actions across the app and the operator panel (see the list below). Every action is a suggestion: you preview it, edit it, then accept it, and nothing is saved without that step. Gemini is the default real provider and it is bring-your-own-key and off until you add your own GEMINI_API_KEY; Google bills you directly and no key, model or credit is included. Résumés are personal data, so every request is strictly de-identified first: names, emails, phone numbers, street addresses and links are removed before anything leaves your server, and keys and card numbers are always stripped. The built-in resume tasks (bullet rewrites, summaries, tailoring, cover letters, skill suggestions, tone check) run on a simulated engine in the demo, clearly marked "Simulated"; connect Gemini (or an OpenAI-compatible or Anthropic key) to get real results. Live calls are tested against recorded fixtures only. The keyword match score is not AI at all: it is a fixed, repeatable calculation against a visible skills list. A Google Translate language switcher sits in the site header and the app top bar.
Gemini AI admin tools, by page: rewrite a summary in the editor, flag gaps and risks in a resume, explain a match report, summarise and classify a job description, polish a cover letter and draft the follow-up email, ask in plain English on My resumes, clean up pasted import text, flag unusual public-link views, suggest next actions for coaches and teams, explain credit usage, flag risky audit activity, and in the operator panel explain the overview, classify skill terms and flag links to moderate.
No live demo: 37 screenshots and two video tours show the marketing site, the editor, tailoring, the match report, sharing, coach review and the operator panel. This product is a web application: there are no mobile apps.
For people who write resumes
- Editor: structured sections (contact, summary, experience, education, skills, projects, certifications, languages, awards, volunteering, custom), live A4 preview, autosave and a version timeline where a restore adds a new version.
- Twelve templates with accent colour, font, text size, spacing and a section manager. PDF export for all twelve and Word (DOCX) export for four of them (ATS Plain, Modern, Compact, Two column).
- Tailor to a job: paste a job description, get proposed changes as a list with before and after, and accept them one by one into a new version.
- Match score: a deterministic keyword score with found and missing terms, from a 321-term starter skills dictionary that you can edit or extend by CSV import.
- Cover letters in a tone you choose, editable, with PDF download. Import from PDF, DOCX, JSON Resume or pasted text, then review the sections it found.
- Share links: a private page per resume with optional password, expiry and download switch, plus view analytics.
For coaches and operators
- Coach workspaces: invite clients, comment on resumes, approve the version you reviewed. By default a coach sees only the clients assigned to them; clients can hide a resume.
- Credits and plans: AI tasks reserve credits, then settle or refund exactly once, with a ledger and a reconciliation command. Four plans and credit packs are included. Payments are simulated until you connect your own gateway (Stripe and Razorpay adapters included).
- Operator panel: overview, AI settings and cost table per task, template catalogue, skills dictionary, link moderation and an audit log.
- REST API with OpenAPI, API keys and webhooks. Setup wizard, demo data, installer lock and a documentation page in the download.
Demo data (real numbers from the seed)
30 workspaces, 107 users, 167 resumes, 391 versions, 146 AI jobs (including failed and refunded ones), 43 share links and 57 invoices. Loading takes about half a minute and a reconciliation command confirms the credit ledger.
Good to know
- Starting codebase for a resume-builder service, not a finished business. Extended licence needed to sell the service to paying customers.
- No model or key included: the demo uses a sample engine. Add your own Gemini key (or another provider's) for real AI text; usage is billed to you by the provider.
- The PDF export does not shape Arabic or other right-to-left scripts (a limit of the PDF library). Latin-script resumes are fully supported; Word export uses Calibri.
- The skills dictionary ships 321 terms, not thousands; extend it from the operator panel. The match score is not an applicant-tracking-system simulation.
- No job-board integrations, no auto-apply, no scraping job posts by URL and no LinkedIn import. Web only.
- Needs MongoDB 6+ (replica set; a single node works) and two long-lived Node processes: the web app and the worker. Back up your
APP_SECRET.
- All people, employers and schools in the demo are fictional.
Hero — Build a clear resume, tailor it to the job and send it as a perfect PDF, with a desk scene and a live match score.
Templates — Twelve layouts with PDF export and Word export for four of them, all from the same content.
