Scholium Documentation#

Scholium — narrated video from markdown slides Scholium — narrated video from markdown slides

Automated instructional video generation from markdown slides with embedded narration.

Scholium (Greek: σχόλιον) — An explanatory note or commentary. Your digital scholium for the modern classroom.

Scholium transforms markdown slides with embedded :::notes::: blocks into professional narrated instructional videos using AI voice synthesis.

Python 3.11+ License: MIT

Quick Start#

Installation:

pip install scholium[piper]

Create a lecture in lecture.md:

---
title: "Python Functions"
title_notes: |
  Welcome to this lesson on Python functions.
---

# What is a Function?

A reusable block of code.

::: notes
Functions are fundamental building blocks in Python.
They let you organize code into reusable pieces.
:::

Generate video:

scholium generate lecture.md output.mp4

Key Features#

Unified Markdown Format

Write slides and narration together using :::notes::: blocks.

Advanced Timing Control

Precise control over slide duration and pauses:

::: notes
[PRE 2s] [POST 3s] [MIN 10s]

Narration with timing directives.
:::
Incremental Reveals

Synchronized bullet-by-bullet reveals:

>- First point
>- Second point
>- Third point

::: notes
Narration for first point.

Narration for second point.

Narration for third point.
:::
Multiple TTS Providers

Choose from eight text-to-speech (TTS) engines — local or cloud, free or commercial:

  • Piper - Fast, local, recommended for beginners

  • ElevenLabs - Highest quality cloud API

  • Coqui - Local voice cloning from audio samples

  • OpenAI - Cloud API with multiple built-in voices

  • Bark - Expressive local synthesis (slow)

  • F5-TTS - Fast zero-shot local voice cloning

  • StyleTTS2 - Expressive diffusion-based local cloning

  • Tortoise - Highest quality local voice cloning (slow)

Indices and tables#