Study Mentor

A voice-first tutor that runs on your own computer. Talk through your lesson, get quizzed, and watch it draw the idea.

Get it on GitHub Quick start
Study Mentor home screen: day plan, study modes, homework list, pixel crab mentor and whiteboard

What it does

Teaches, does not read

Explains in its own words, asks questions, and pushes back. It engages you instead of reciting text.

Draws as it explains

Tables, key points, and Mermaid diagrams with each node highlighted step by step. The whiteboard resizes itself to fit the content.

Remembers what you miss

Local SQLite progress tracking. Weak topics come back in future sessions, and your streak is preserved.

Climbs a mastery ladder

Every skill moves from taught to recognised, applied, exam-style and professional. Misses drop you a rung and re-teach; skills come back on a spaced schedule.

Grades honestly

An independent grader scores your answers in code, so praise is earned. Practice items with named distractor traps are graded by the server, not the model.

Carries everything forward

Skipped sections and mistakes roll into the next session, and a day only completes after the material is covered. Homework stays on your list until you tick it.

Private and cheap

Speech runs on your machine. Only text goes to the language model API you choose. About a few dollars a month with a cheap model.

How it works

Quick start

  1. Clone the repository and enter the directory.
  2. Create a virtual environment and install dependencies.
  3. Add your API key to ~/.config/study-mentor/deepseek.key.
  4. Download TTS models into ~/.local/share/study-mentor/tts.
  5. Run the server and open the local URL.
git clone https://github.com/consultantstanleymn/study-mentor.git
cd study-mentor
python -m venv venv
venv/bin/pip install fastapi "uvicorn[standard]" httpx beautifulsoup4 lxml faster-whisper numpy soundfile kokoro-onnx python-multipart
put your API key in ~/.config/study-mentor/deepseek.key
download kokoro-v1.0.onnx and voices-v1.0.bin into ~/.local/share/study-mentor/tts
./run.sh
open http://127.0.0.1:8765

Bring your own curriculum

Day-based folders (days.json plus day pages) or a week-based checklist file. Register your content in server/content.py.

Keys and shortcuts