rachid chabane.
Search
← All projects Project · 03 · agent-maintained

CCA-F Exam Trainer: Bilingual Certification Practice

in production 2026 React 19 · TypeScript · Vite FR / EN
300+ scenario questions
5 weighted domains
11 course summaries
EN / FR fully bilingual

01What it is

A fully client-side, bilingual (English and French) practice trainer for Anthropic's Claude Certified Architect Foundations (CCA-F) exam: timed mocks weighted across the five official domains, a scaled score with a per-domain breakdown and a full answer review, plus a study mode of original course summaries grounded in first-party Anthropic documentation.

02Engineering

CCA-F Exam Trainer is a fully client-side single-page app (Vite, React 19, TypeScript, Tailwind CSS v4) with no backend, no database, and no accounts: every in-progress exam, results history, and preference lives only in the browser’s localStorage. The exam runner assembles a timed mock weighted across the five official domains, with a flag-and-navigator grid, a 120-minute countdown, and auto-submit at zero. Each sitting is graded on a documented linear approximation of the scaled 100-1000 score (pass at 720), with a per-domain accuracy breakdown and a full answer review that shows your answer, the correct one, why it is best, and why each distractor falls short. A Zustand store persists session state, so an in-progress mock survives a refresh or a slept laptop, recent attempts build a score history, and you can re-quiz only the questions you missed or drill a single weak domain untimed. The English and French toggle is total: one click swaps every label and every piece of content, including stems, options, explanations, and the study-mode course bodies. All questions and summaries are original and grounded in first-party Anthropic documentation, with no third-party question banks; a deterministic fact-check pins the exam parameters and keeps community-reported numbers labeled as such, an optional Claude-powered pass reviews the answer keys, and a Playwright suite runs against the real production bundle. Continuous integration ties it together: every push runs schema validation, the fact-check, and end-to-end tests before deploying to GitHub Pages.

03 Architecture

Content (bilingual JSON)
question poolcourse summariesexam blueprint
App (Vite, React 19, TS)
exam runnerscaled scoringstudy readerEN / FR + theme
State
Zustand storelocalStorage (resume, history, prefs)
Quality gates (CI)
schema checkdeterministic fact-checkPlaywright e2e
Deploy
GitHub ActionsGitHub Pages
Local-first, content-validated, continuously deployed

04 Highlights

  • A timed mock built like the real exam: weighted across the five official domains, with a flag-and-navigator grid, a 120-minute countdown, auto-submit, and a scaled 100-1000 score (pass at 720).
  • Every result breaks accuracy down per domain and opens a full review: your answer, the correct answer, why it is best, and why each distractor falls short.
  • A total English / French toggle flips all UI and all content at once; language, theme, and an in-progress attempt persist locally, with no backend and no accounts.
  • Original content grounded in first-party Anthropic docs, guarded by a deterministic fact-check, an optional Claude answer-key review, and a Playwright suite that runs against the production bundle before each GitHub Pages deploy.

05Stack

React 19TypeScriptViteTailwind CSS v4ZustandPlaywrightGitHub Actions

06Status

in production

07 Links

08 Related articles