Skip to content
Truth Academy.
Explore Feynman
Truth Academy
Tutor setup

Run the tutor on your computer

Feynman's tutor uses a free AI model that runs on your own machine, so nothing you type is sent anywhere. The tutor is off on this website because there is no model here for it to use. Set it up once, in about 15 minutes (mostly download time), and the full tutor works on your computer.

What you need

  • A Mac, Windows, or Linux computer with at least 8 GB of memory
  • About 3 GB of free disk space
  • Node.js 18 or newer (the LTS download from nodejs.org)
  • Access to the Feynman code (see step 3)
  1. 1

    Install Ollama

    Ollama is a free app that runs AI models locally. Download it from ollama.com/download and open it once. On a Mac or Windows it then keeps running in the background. On Linux, install it from a terminal:

    curl -fsSL https://ollama.com/install.sh | sh
  2. 2

    Download the tutor's model

    In a terminal, run this (about 1 GB):

    ollama pull qwen2.5:1.5b

    Check that it works. It should reply with a short greeting:

    ollama run qwen2.5:1.5b "Say hi"
  3. 3

    Get Feynman

    Download the code and its dependencies:

    git clone https://truth-computing-incorporated.ghe.com/truth-computing/truth-academy.git
cd truth-academy/learn-feynman-main
npm install

    The code lives in Truth Computing's repository. If you don't have access yet, ask us through school partnerships.

  4. 4

    Start it

    npm run dev

    Open http://localhost:3210, pick any classroom module, and ask the tutor a question.

Optional: a stronger model

A larger model follows the tutor's rules (questions first, no doing homework for you) more reliably. It needs about 5 GB of disk and works best with 16 GB of memory.

ollama pull qwen2.5:7b-instruct

Then create a file named .env.local in the learn-feynman-main folder containing this line, and restart npm run dev:

FEYNMAN_TUTOR_MODEL=qwen2.5:7b-instruct

If something goes wrong

“The tutor is temporarily unavailable.”
Ollama isn't running. Open the Ollama app, or run ollama serve in a terminal, then ask again.
“The tutor model is unavailable right now.”
The model isn't downloaded. Run ollama pull qwen2.5:1.5b (or the model named in your .env.local).
“The AI tutor is turned off for this deployment.”
You're on the website rather than your own copy. Open http://localhost:3210 instead.
Port 3210 is already in use.
Start it on another port with npx next dev -p 3211, then open http://localhost:3211.