MentorVerse is a VS Code extension that gives you a team of senior specialist AI mentors who teach you interactively — from basic → intermediate → advanced — across DSA, programming languages, DevOps/containers, cloud, and networking. It runs on your existing GitHub Copilot subscription via the VS Code Language Model API (no API keys, no extra cost).
Meet the mentors
| Mentor | Role |
|---|---|
| 🧭 Ada | Lead Mentor & Coach — coordinates, motivates, tracks your progress |
| 💡 Grace | Senior Developer & DSA — algorithms, data structures, logic-building |
| 🏛️ Archie | Software Architect — design, patterns, trade-offs |
| 🚢 Kai | Senior DevOps Engineer — Docker, Kubernetes, CI/CD |
| 🛰️ Nova | Senior Network Engineer — protocols, topologies, troubleshooting |
| 🗄️ Cody | Database & SQL Specialist — queries, optimization, schema design, modern data |
| 📊 Flo | Data Scientist — statistics, pandas/numpy, EDA, experiments |
| 🤖 Alan | AI/ML Specialist — algorithms, training, evaluation, deep-learning intro |
| 🧠 Lex | LLM & GenAI Specialist — transformers, prompting, RAG, agents |
Ask Ada anything; she pulls in the right specialist on demand.
Features (rolling out by phase)
- Multi-persona mentor chat with a coordinating lead mentor.
- Four learning tracks: DSA + Programming, DevOps (Docker/K8s), Cloud (Azure/AWS/GCP), Networking.
- Leveled curriculum (basic → advanced) with on-demand, cached lessons.
- Hands-on local code playground (Python / JavaScript / C#) with auto-graded challenges.
- Hands-on SQL playground (Node's built-in SQLite) with a free-play scratchpad and auto-graded query challenges that compare your result set to a reference.
- Hands-on stats/ML coding challenges (implement mean, variance, distance, cosine similarity, accuracy, gini, entropy… from scratch in Python; numpy/pandas allowed) and an interactive Prompt Lab to experiment with LLM techniques (zero/few-shot, system prompts, chain-of-thought, mini-RAG) using your Copilot model.
- Endless practice: Grace generates fresh challenges on demand, auto-verified by running her reference solution against the tests before they're shown; verified ones are saved.
- Online playground launchers for Docker/Kubernetes (Killercoda / Play-with-Docker).
- Gamification: XP, levels, streaks, badges, and a progress dashboard.
- Read-aloud (text-to-speech): listen to lessons and mentor replies via the built-in Web Speech API — with a voice/speed/pitch settings panel and an optional auto-read toggle (free, offline, no keys).
- "Explain simply / real-world analogy" mode and struggle detection.
Develop & run
npm install
npm run build # one-shot build (host + webview)
npm run watch # rebuild on change
Then press F5 in VS Code to launch the Extension Development Host, and run the command “MentorVerse: Open Mentor Hub” from the Command Palette (or click 🎓 MentorVerse in the status bar).
Install the packaged build
npm run package # produces mentorverse.vsix
code --install-extension mentorverse.vsix --force
Reload VS Code, then open the hub from the Command Palette or status bar. The mentors use your GitHub Copilot subscription via the VS Code Language Model API — the first request will ask you to allow model access.
Configuration (Settings → "MentorVerse")
mentorverse.preferredModelFamily— Copilot model family the mentors request (defaultgpt-4o).mentorverse.defaultLanguage— default code language for the playground (python/javascript/csharp).mentorverse.dailyGoalXp— daily XP goal for streaks (default100).mentorverse.disabledMentors— list of mentor IDs to turn off (hidden from chat & routing). The lead mentoradais always on. IDs:grace,archie,kai,nova,cody,flo,alan,lex. Example:"mentorverse.disabledMentors": ["cody", "nova"].mentorverse.hideDisabledMentorTracks— whentrue, a disabled mentor's learning track and challenges are also hidden (not just chat). Changes apply live (no reload needed).
Tech
- TypeScript + esbuild (extension host & webview bundles)
- React webview UI
- VS Code Language Model API (Copilot) for the mentors
- Local JSON persistence in the extension's global storage
Status: under active development. See
plan.md(session) for the phased roadmap.
Created by Vishal Agarwal. Licensed under the MIT License (see the LICENSE file).