---
name: explain
description: Create a Scrimba video explainer — an instantly generated video with narration, animations, mermaid diagrams, LaTeX maths, and precisely composed code walkthroughs, rendered live on scrimba.com. Use when the user wants something explained visually or as a video, or asks for an explainer or walkthrough of code, a concept, a flow, an architecture, a git branch, a PR, or a diff. Also use for playlists or courses of lessons.
---

# Scrimba Explain

Scrimba Explain turns anything into a rich explainer with narration and visual aids, built live at a shareable scrimba.com URL. It shines for:

- Code diffs (side-by-side before/after)
- Concepts and learning topics
- Flows and architectures
- Code walkthroughs with precise highlighting
- Mermaid diagrams and LaTeX maths

## How to use it

The `scrimba-explain` MCP server (bundled with this plugin) provides the tools:

1. **One focused explainer**: call `start_explainer_stream`, then push content with `append_explainer_chunk`, and close with `finish_explainer_stream`. Scrimba renders the OPML you push live.
2. **A playlist or course of lessons**: use `create_playlist` — you plan the curriculum and author every lesson through the same OPML streaming flow.

## The `start_explainer_stream` result IS the authoring contract (~50 KB)

The return value of `start_explainer_stream` is not a status message. It contains the stream URL plus the **complete OPML authoring contract** — the format specification you must follow in every subsequent `append_explainer_chunk` call. Handle it accordingly:

- **It is ~50 KB.** Many agent harnesses truncate or elide tool results (limits like 2,000 bytes or 10,000 characters are common). Before calling the tool, make sure you can receive and retain a payload of that size.
- **If your environment truncates large tool results**, use whatever file/storage tools you have to save the full result to a file, then read that file to the end before pushing any chunk.
- **Never author chunks from a partial contract.** If you only saw a truncated result and cannot recover the rest, say so rather than guessing at the format.

## Important workflow notes

- As soon as `start_explainer_stream` returns the URL, give it to the user — they open it to watch the explainer get built live.
- The streaming and tokens are internal mechanics for you; don't narrate them to the user. Just say you're creating their Scrimba explainer.