Learn it.Build with it. Curricula, graded exercises, past exam papers and practical tools for software, languages and career work. Read for free; practise with real feedback.
Real 3D playground Drive into CuongThai
This is a static preview. The 3D engine and its assets only load after you choose to enter.
Explore in 3D Long reads worth finishing One topic at a time, explained far enough that you can leave with something working.
01 How to Use the Command Line in Linux and macOS Navigate, move files, edit text and chain commands — from zero.
Read guide 02 The Event Loop, Callbacks, Promises and Async/Await Why JavaScript never blocks, and where your await actually goes.
Read guide 03 Node.js From Zero to Production 19 chapters: runtime, HTTP, PostgreSQL, auth, queues, observability.
Full course · 112 lessons Open resource 04 An Introduction to GraphQL Schemas, resolvers and why it is not a REST replacement.
Read guide 05 How to Structure and Organize a React Application Folders, state boundaries and the seams that survive a rewrite.
Read guide 06 A Complete Guide to CSS Concepts and Fundamentals The box model, cascade, specificity, flex and grid — properly.
Read guide 07 How to Use Redux and React Store, actions, reducers — and when you genuinely need them.
Read guide 08 How to Use Vue, the JavaScript Framework Reactivity, components and the single-file component model.
Read guide 09 How to Set Up webpack From Scratch Entry, output, loaders, plugins — a bundler you can reason about.
Read guide 10 Shell Scripting for People Who Deploy Things set -euo pipefail, traps, and why your script eats its own errors.
Read guide 11 How to Set up a Mac for Development Shell, Homebrew, Node, Docker, keys — a machine ready to work.
Read guide 12 Reading Production: Logs, Metrics and a Calm Head What to look at first when the thing you shipped is on fire.
Read guide 13 Building a Jira-style Project Tool (CT Work) How CT Work was built in 21 days: the code, the bugs we fixed, the timeline — and how to use it well.
Bài đọc · Tiếng Việt · tải app ở cuối bài Read guide 14 AI Code on the Desktop: An Agent That Codes With You Inside the coding agent of the desktop app: the tool loop, safety, models by measured cost, bugs fixed.
Bài đọc · Tiếng Việt · tải app ở cuối bài Read guide Flagship tools
Two tools we built and use every day Not just lessons — real software with real users: students, lecturers and mentors, and teams at software companies. A Jira-style place to run projects, and a coding agent that runs on your own machine — and the two plug into each other.
0:06 My work0:10 Sprint & board0:21 SRS & diagrams0:30 Testing0:35 Ask AI0:43 Desktop app0:49 Agents & MCP
CT Work A Jira-style workspace built around how Vietnamese students, lecturers and software teams actually work — boards and sprints, SRS documents, FPT-format test reports, classrooms, and AI agents that pick up cards with you.
Made for
Student capstone teams Lecturers & mentors Company project teams Freelancers with clients
screens 68
agent commands (MCP) 123
places: web · app · standalone 3 Features Plan & track Drag-and-drop board, backlog, list Sprints, timeline (Gantt), releases JQL search and saved filters Timers, worklogs, workload Capstones & classrooms SWP391 · SWR302 · SWT301 · SEP490 templates Lecturer hub, classes, submission portal Per-member contribution, weekly reports Defense practice, AI requirement review Requirements & docs Structured SRS: use cases, actors, rules Traceability matrix (RTM), Wiegers set Real-time collaborative docs Diagram Studio: diagrams from your data Testing & quality Xray-style test cases and cycles FPT-format unit/integration reports Reviews, inspections, baselines Spec-fidelity scoring for cards AI in the project Ask AI about the open project (⌘J) AI agents as real members with an owner Connect Claude Code over MCP in one command Agent “done” lands in Review for a human Collaboration Project chat channels and threads Meetings, minutes, action items Planning poker, retros, OKRs Real-time notifications, .ics calendar Client delivery Client portal, UAT, prototype review Change requests, RAID, approvals Service desk with SLAs Project finance, presentation mode Integrations & import Microsoft 365 · Google · Notion · Slack · GitLab Import from Trello · Asana · Jira · CSV Forms → cards, external channels → intake API tokens, project export/import, 2FA See all 8 feature groups Why it stands out 01 Speaks the FPT capstone language Project templates, SRS sets and 5.1/5.2/5.3 test reports in the school’s own formats — not a Jira workaround.
02 AI that works like a teammate Agents have their own account and an accountable owner, and the server blocks them from deleting, approving or contacting clients.
03 Everywhere, one dataset Web, the CuongThai app and the standalone CT Work app — edits sync live for the whole team.
04 Bilingual, free for students Each member picks Vietnamese or English; AI-written text follows the project language.
0:04 The task0:09 Reading0:17 Plan0:21 Permission0:25 Tests0:31 Self-check0:36 Result0:52 Models
AI Code A coding agent that runs on your own machine: it reads the whole project, edits many files, runs commands in a real terminal, opens a browser to check its own work — and stops to ask before anything risky.
Made for
Students doing labs & capstones Developers on real projects Beginners who want each step explained Anyone who wants an agent without a CLI
agent tools 46
models to pick per task 8
operating systems 3 Features Read & edit code Reads the whole project with glob/grep Multi-spot, multi-file edits in one pass Review the diff before accepting Reads PDF, Word/Excel, notebooks, images, zip Run & verify A real terminal (PTY), password prompts work Background commands: dev servers alongside Your real shell PATH — node, docker, psql Reads failures and fixes until green A browser for the agent Opens pages, clicks and types like a user Screenshots and console logs Downloads docs, reads many pages at once Checks the UI it just changed Git & hand-off Controlled git status/diff/commit Opens pull requests Isolated worktrees for experiments Writes results straight into Notes Extend it Plug in MCP servers (CT Work included) A reusable skills library Sub-agents for big tasks Project memory across sessions Safety & cost 5 permission modes: plan → ask → auto-edit Blocks destructive commands and secret dirs Daily AI wallet with cost caps Busy gateway ⇒ automatic switch to an equivalent model See all 6 feature groups Why it stands out 01 Runs on your machine, with your code Nothing is uploaded; the agent works in the folder you choose, with the tools you already have.
02 Models picked by measurement The default is the cheapest model inside the tool loop — measured 7× cheaper than the “premium” one for the same job.
03 Checks before it says done Runs tests, opens the browser, reads the console — “done” comes with evidence, not hope.
04 Wired into CT Work Picks up a card, does it, logs work and moves it to Review over MCP — you still approve.
01
Learn
A clear order, from the first lesson Full curricula instead of scattered tutorials — each one has an order, exercises and exams.
6 ways in 02
Build
Tools that turn learning into work A Jira-style workspace, a desktop AI coding app, notes, diagrams and a CV that gets read — the things that turn study into output.
6 ways in 03
Practise & play
Stop reading. Put it under pressure. Rehearse in realistic conditions, then take a break somewhere that has nothing to do with work.
6 ways in 
Projects · capstones · classrooms