AI Coding Teacher
An interactive Streamlit app that acts as an AI coding mentor. It adapts to user skill level, analyzes questions, streams step-by-step explanations with examples, auto-continues long answers. Powered by the GPT-5 model through the AIML API.
Core trick
Streamlit app auto-continues replies that hit the token limit and summarizes older turns to preserve context.
Description
$26
README excerpt
# AI Coding Teacher An interactive Streamlit app that acts as a patient, Socratic AI mentor for learning programming. It adapts depth by level, explains pasted code, suggests improvements, generates exercises, tracks topical progress, summarizes long chats, streams responses, and auto‑continues when truncated. --- ## 🚀 Features - Teaching modes: Beginner / Intermediate / Advanced (adaptive tone & depth) - Code detection → explanation, issues, refactors, practice task - Periodic conversation summarization to preserve context window - Streaming responses with live typing effect - Auto continuation when reply hits token limit - Adjustable generation controls (temperature, top‑p, max output tokens, max continuations) - Lightweight topic progress tracking - Token budget estimation (optional `tiktoken`) - Download full chat history as JSON - Safe context management with buffer --- ## 🧱 Tech Stack | Layer | Tools | |-------|-------| | UI | Streamlit | | Model API | OpenAI‑compatible endpoint (`openai` Python client) | | Language | Python 3.10+ | | Optional | `tiktoken` for token counting | --- ## 📁 Project Structure ``` app.py Main Streamlit application utils.py Chat history init + export helpers requirements.txt Dependencies progress.json Created at runtime (topics) screenshots/ (Add your images) ``` --- ## 🔑 Environment Configuration To run this app, you need an API key from "aimlapi.com". Set your API key (choose one method): Shell: ```bash export