Orpheus

The Orpheus System transforms static slides into interactive lecture videos with lifelike professor avatars, combining expressive narration, visual presence, and dynamic content to create engaging, personalized learning experiences.

Table of Contents

• Architecture
  • Service Components
  • API Interface Documentation
• Getting Started
  • User Guide
    • Lecturer View
    • Student View
  • Development Setup
    • 1. Install Python 3.13.7 using pyenv
      • Linux (Debian/Ubuntu)
      • macOS
      • Windows (PowerShell, run as Administrator)
    • 2. Install Poetry 2.2.1
      • Linux / macOS
      • Windows (PowerShell)
  • Deployment

Architecture

Service Components

The Orpheus system is composed of multiple specialized services, each handling different aspects of the lecture generation pipeline. Below you can find links to the technology-specific documentation and implementation details for each service component.

API Interface Documentation

Getting Started

User Guide

Lecturer View

The lecturer view can be accessed via the Admin Button located at the top right.

To personalize the course delivery, the lecturer is required to upload both avatar images and voice samples, followed by the relevant course materials.

---

👤 Avatar Uploads

Three distinct avatars should be provided to represent different stages of the lecture:

• Beginning Avatar

  • Used at the beginning of the lecture.
  • Recommended: a happy facial expression to create a welcoming atmosphere.

• Default/Middle Avatar

  • Used during the main lecture delivery.
  • Recommended: a neutral facial expression to maintain focus.

• Ending Avatar

  • Used at the end of the lecture.
  • Recommended: a happy facial expression to close on a positive note.

---

🎙️ Voice Samples

Similarly, three voice samples should be uploaded, aligned with the same lecture stages as the avatars:

• Beginning Voice Sample — welcoming and engaging.
• Default/Middle Voice Sample — clear and neutral delivery.
• Ending Voice Sample — positive and encouraging tone.

---

📑 Course Materials

• Upload course slides and/or pre-recorded lecture videos.
• These materials will be processed and integrated into the system by the Document Intelligence Team.
• For further details, refer to the Document Intelligence README.

---

Student View

1. Choose your level of expertise by selecting a suitable character:

2. Enter your question or choose from the predefined ones:

3. Wait for for the generation process. In the meantime a textual answer will be given. The lectre will start as soon as the first video is done:

4. Watch the video:

Development Setup

Environment Variables

1. Add a .env file in the root directory

   cp exampleEnv .env
   

2. Make sure to supply values for at least one AI-Model (e.g. AWS) and define the respective model names (MODEL_NAME, SPLITTING_MODEL, SLIDESGEN_MODEL)
   1. How to get AWS Keys? (with Hackathon license)
      1. You need an AWS Hackathon Account
      2. Login at https://slalom-hackathon.awsapps.com/start/#/?tab=accounts
      3. Go to slalom_IsbUsersPS
      4. Go to Amazon Bedrock
      5. View API keys
      6. Click Generate short-term API keys (Generate long-term API keys did not work with our license)
      7. Copy the upper API Key that starts with bedrock-api-key...
   2. Update the model (AWS includes some user specific tag there, so API keys of other people won't work)
      1. You need an AWS Hackathon Account
      2. Login at https://slalom-hackathon.awsapps.com/start/#/?tab=accounts
      3. Go to slalom_IsbUsersPS
      4. Go to Amazon Bedrock
      5. Infer -> Cross-region inference
      6. Choose the model that you want to use, copy the Inference profile ARN that starts with arn:aws:bedrock:...
3. You can overwrite the global .env file values with service specific .env files

1. Install Python 3.13.7 using pyenv

The project is based on python 3.13.7. We recommend using pyenv to manage your python versions.

Linux (Debian/Ubuntu)

1. Install system dependencies for building Python (one-time setup)

   sudo apt update
   sudo apt install -y build-essential curl git \
   libssl-dev zlib1g-dev libncurses5-dev libbz2-dev libreadline-dev \
   libsqlite3-dev libffi-dev liblzma-dev tk-dev uuid-dev
   

2. Install pyenv (via the pyenv-installer)

   curl https://pyenv.run | bash
   

3. Add pyenv to your shell and reload it
   1. For bash

      echo 'export PYENV_ROOT="$HOME/.pyenv"' >> ~/.bashrc
      echo 'command -v pyenv >/dev/null || export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.bashrc
      echo 'eval "$(pyenv init -)"' >> ~/.bashrc
      

   2. For zsh

      echo 'export PYENV_ROOT="$HOME/.pyenv"' >> ~/.zshrc
      echo 'command -v pyenv >/dev/null || export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.zshrc
      echo 'eval "$(pyenv init -)"' >> ~/.zshrc
      

   Important: Restart your terminal or run exec "$SHELL" for the changes to take effect

   exec "$SHELL"
   

4. Install Python 3.13.7 and set it as the global default version

pyenv install 3.13.7
pyenv global 3.13.7

5. Verify the installation

python --version

Expected output: Python 3.13.7

macOS

1. Install Homebrew if you haven't already

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
   

2. Install pyenv

   brew install pyenv
   

3. Add pyenv to your shell and reload it For zsh (the default on modern macOS):

   echo 'export PYENV_ROOT="$HOME/.pyenv"' >> ~/.zshrc
   echo 'command -v pyenv >/dev/null || export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.zshrc
   echo 'eval "$(pyenv init -)"' >> ~/.zshrc
   

   Important: Restart your terminal or run exec "$SHELL" for the changes to take effect

   exec "$SHELL"
   

4. Install Python 3.13.7 and set it as the global default version

   pyenv install 3.13.7
   pyenv global 3.13.7
   

5. Verify the installation

   python --version
   

   Expected output: Python 3.13.7

Windows (PowerShell, run as Administrator)

1. Install pyenv-win via PowerShell

   Invoke-WebRequest -UseBasicParsing -Uri "https://raw.githubusercontent.com/pyenv-win/pyenv-win/master/pyenv-win/install-pyenv-win.ps1" -OutFile "./install-pyenv-win.ps1"; &"./install-pyenv-win.ps1"; Remove-Item "./install-pyenv-win.ps1"
   

Troubleshooting Common Installation Issues

1. Script Execution is Disabled

• Issue: You receive an error in PowerShell stating ...cannot be loaded because running scripts is disabled on this system.
• What to do: This is due to PowerShell's Execution Policy. Run PowerShell as Administrator and execute the following command to allow the script to run for the current session, then try the installation command again.

  Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope Process
  

2. pyenv Command Not Found After Installation

• Issue: After the installer finishes, opening a new terminal and typing pyenv results in a command not found error.
• What to do: The installer couldn't modify your User PATH environment variable correctly, or your terminal session needs to be refreshed.
  1. Restart your terminal: Close and reopen PowerShell/CMD completely.
  2. Restart your computer: A full restart will ensure environment variables are reloaded.
  3. Manually add to PATH: If it still fails, you must add the following two paths to your User PATH environment variables.
     • %USERPROFILE%\.pyenv\pyenv-win\bin
     • %USERPROFILE%\.pyenv\pyenv-win\shims

3. System Python Overrides pyenv Version

• Issue: You've set a Python version with pyenv global or pyenv local, but running python --version shows your old system version (or opens the Microsoft Store).
• What to do: This is a PATH priority issue.
  1. Disable Windows App Execution Aliases: Go to Start > Manage App Execution Aliases and turn off the aliases for python.exe and python3.exe. This is the most common cause.
  2. Check your PATH order: Ensure the pyenv shims and bin paths appear before any other Python installation paths in your environment variables.

4. Shims Are Not Working for New Packages

• Issue: You install a package with a command-line tool (like pipx or poetry) using pip, but the command isn't available in your terminal.
• What to do: You need to rebuild the pyenv shims so it's aware of the new executable.

  pyenv rehash
  

2. Add pyenv to your PowerShell session The following lines are automatically added to your $PROFILE. You may need to run them manually for the current session or restart your terminal. ```powershell $env:PYENV = [System.Environment]::GetEnvironmentVariable('PYENV','User') $env:PYENV_HOME = [System.Environment]::GetEnvironmentVariable('PYENV_HOME','User') $env:PYENV_ROOT = [System.Environment]::GetEnvironmentVariable('PYENV_ROOT','User') $env:Path = [System.Environment]::GetEnvironmentVariable('path','User') ``` **Important**: Restart your PowerShell window to ensure the PATH changes are active. 3. Update pyenv-win to get the latest list of available versions ```powershell pyenv update ``` 4. Install Python 3.13.7 and set it as the global default version ```powershell pyenv install 3.13.7 pyenv global 3.13.7 ``` 5. Verify the installation ```powershell python --version ``` Expected output: Python 3.13.7

(Optional) Check which version pyenv is managing

pyenv version

Expected output: 3.13.7 (set by C:\Users\YourUser.pyenv\pyenv-win\version)

2. Install Poetry 2.2.1

We use Poetry as our dependency and environment management tool.

Linux / macOS

1. Install Poetry using the official installer script

   curl -sSL https://install.python-poetry.org | python3 -
   

2. Update Poetry to version 2.2.1

   ~/.local/bin/poetry self update 2.2.1
   

3. Ensure Poetry is on PATH for the current shell session

   export PATH="$HOME/.local/bin:$PATH"
   

   1. Persist Poetry on PATH for bash

      echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bashrc
      

   2. Persist Poetry on PATH for zsh

      echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc
      

4. Verify Poetry installation

   poetry --version
   

Windows (PowerShell)

1. Install Poetry using the official installer script

   (Invoke-WebRequest -Uri https://install.python-poetry.org -UseBasicParsing).Content | python -
   

   Poetry is typically installed under:

   $env:APPDATA\Python\Scripts
   

2. Update Poetry to version 2.2.1

   poetry self update 2.2.1
   

3. Ensure Poetry is on PATH for the current PowerShell session

   $env:Path += ";$env:APPDATA\Python\Scripts"
   

4. Persist Poetry on PATH for future sessions

   setx PATH "$($env:PATH)"
   

5. Verify Poetry installation

   poetry --version
   

Deployment

Run this command in your root directory to build all services, start them up under the project name orpheus, and pull any necessary images:

docker compose -p orpheus up --build