⚡ Accelode — Escape Tutorial Hell

Stop watching tutorials. Start building.

Accelode analyzes your GitHub profile, identifies your skill gaps through an adaptive quiz, and generates personalized micro-projects with an AI mentor that teaches you concepts — not just code.

Features

1. GitHub Profile Analysis

Paste your GitHub URL — Accelode fetches your real repos, languages, stars, and profile README to understand your current skill level.

2. Adaptive Diagnostic Quiz

A 10-question quiz that adapts in real-time. Each question is generated based on your previous answers to pinpoint exactly where your knowledge gaps are.

Every question includes a "Not sure / Teach me" option — no shame in learning.

3. Personalized Learning Topics

After the quiz, AI generates 8 tailored topics ordered from beginner to advanced. Each topic includes:

Description of why it matters
Suggested articles to read
A hands-on project with step-by-step syllabus

4. AI Mentor Chat (Socratic Method)

Select a topic and start chatting with your AI mentor. Unlike code generators, Accelode's mentor:

Teaches concepts first — uses real-world analogies
Never gives code unless you explicitly ask
Checks your understanding before moving forward
Responds in your language (English, Azerbaijani, etc.)

5. Code Review

Paste your code in the review panel and get architectural feedback aligned to your current learning topic.

6. Project Scaffold Generation

Generate complete project files with strategic // TODO: [AI HINT] comments directly in your workspace. Includes README, config files, and source code.

7. CodeLens: AI Hints on TODO Comments

Accelode adds inline buttons on every // TODO: comment in your code:

"💡 Get AI Hint" — a pedagogical hint to guide you
"📖 Explain Concept" — architectural context for why this code matters

Works across all programming languages.

8. Offline Mode

No API key? No problem. Accelode works offline with pre-built project templates for Go, Node.js, React, Docker, and Python.

How Is This Different from Copilot?

	GitHub Copilot	Accelode
Goal	Write code faster	Understand concepts deeply
Approach	Auto-generates code	Teaches with analogies, then code
CodeLens	Generate → ready code	Hint → you write the code
Structure	Free-form Q&A	Profile → Quiz → Topics → Mentor
Result	More productivity	More understanding

Use both together: Learn with Accelode, build faster with Copilot.

Getting Started

Install the extension from VS Code Marketplace
Click the ⚡ Accelode icon in the Activity Bar
Paste your GitHub URL and click Analyze
Take the adaptive quiz (or skip it)
Pick a topic and start learning with your AI mentor

Optional: Configure AI

Create a .env file in your workspace root:

LLM_API_KEY=sk-your-openai-key
LLM_BASE_URL=https://api.openai.com/v1
LLM_MODEL=gpt-4o-mini

Supports any OpenAI-compatible API (OpenAI, DeepSeek, Azure, local models).

Without an API key, the extension works in offline mode with mock project templates.

Supported LLM Providers

Provider	Base URL	Model
OpenAI	`https://api.openai.com/v1`	`gpt-4o-mini`
DeepSeek	`https://api.deepseek.com/v1`	`deepseek-chat`
Azure OpenAI	Your Azure endpoint	Your deployment name
Any OpenAI-compatible	Custom URL	Custom model

Extension Commands

Command	Description
`Accelode: Get AI Hint`	Pedagogical hint for a TODO comment
`Accelode: Explain Concept`	Architectural context for a TODO comment

Privacy & Security

No data collection — everything runs client-side in your VS Code
GitHub API — only fetches public profile data
LLM calls — go through your own API key, not our servers
CSP — Content Security Policy with nonce-based script loading
Zero runtime dependencies — only Node.js built-in modules

Requirements

VS Code 1.85.0 or later (including Cursor)
Optional: An OpenAI-compatible API key for AI features

Release Notes

See CHANGELOG for full release history.

1.0.1

Fixed .env loading — API key is now read from ~/.accelode/.env globally, works from any workspace.

1.0.0

Initial release with profile analysis, adaptive quiz, personalized topics, AI mentor chat, code review, project generation, and CodeLens integration.

License

MIT