Skip to content
Skills Library
Education · Building & coding

explainer-video

/explainer-video

Make a short narrated explainer video (3Blue1Brown style) on the user's Mac: Manim animates, the tts command (Kokoro, local) narrates, and two checks catch dropped sound and labels sitting on lines before anyone sees it. Use whenever the user wants an idea, derivation, algorithm or result explained as a video or animation: for a course, a talk, a student, or to understand something themselves. Triggers: 'make an explainer video', '3b1b style', 'animate this derivation', 'Manim', 'a video explaining X', 'narrated animation', 'show this as a video', 'explainer on'. No API key needed. Boundaries:

Licence
MIT
Last repository change
22.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
6
Scripts
4
HTML comments
0
URLs in SKILL.md
1
Deprecation

No signals detected. Repository changed on 22.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What to check before installing
  • 4 scripts in this package. Review them before installing, as Claude may execute them.

Scripts: examples/zigzag/zigzag.py, scripts/check_video.py, scripts/narrate.py, scripts/narrated_scene.py

What it does and when to use it

Make a short narrated explainer video (3Blue1Brown style) on the user's Mac: Manim animates, the tts command (Kokoro, local) narrates, and two checks catch dropped sound and labels sitting on lines before anyone sees it. Use whenever the user wants an idea, derivation, algorithm or result explained as a video or animation: for a course, a talk, a student, or to understand something themselves. Triggers: 'make an explainer video', '3b1b style', 'animate this derivation', 'Manim', 'a video explaining X', 'narrated animation', 'show this as a video', 'explainer on'. No API key needed. Boundaries:

Triggers from the description

make an explainer video3b1b styleanimate this derivationManima video explaining Xnarrated animationshow this as a videoexplainer on
Original author description

Make a short narrated explainer video (3Blue1Brown style) on the user's Mac: Manim animates, the tts command (Kokoro, local) narrates, and two checks catch dropped sound and labels sitting on lines before anyone sees it. Use whenever the user wants an idea, derivation, algorithm or result explained as a video or animation: for a course, a talk, a student, or to understand something themselves. Triggers: 'make an explainer video', '3b1b style', 'animate this derivation', 'Manim', 'a video explaining X', 'narrated animation', 'show this as a video', 'explainer on'. No API key needed. Boundaries:

How it works

  1. Setup
  2. Before animating: settle what it says
  3. Pipeline
  4. Layout that avoids the usual faults
  5. Known failure modes
Original SKILL.md headings · 103 lines, 847 words
SKILL.md preview
---
name: explainer-video
description: "Make a short narrated explainer video (3Blue1Brown style) on the user's Mac: Manim animates, the tts command (Kokoro, local) narrates, and two checks catch dropped sound and labels sitting on lines before anyone sees it. Use whenever the user wants an idea, derivation, algorithm or result explained as a video or animation: for a course, a talk, a student, or to understand something themselves. Triggers: 'make an explainer video', '3b1b style', 'animate this derivation', 'Manim', 'a video explaining X', 'narrated animation', 'show this as a video', 'explainer on'. No API key needed. Boundaries: a static figure is matplotlib; slides are polished-docs-and-decks; audio alone is text-to-audio; a paper read aloud is paper-to-audio."
---

# Explainer video

A narrated Manim animation, made and checked on the user's Mac. Nothing about the video leaves
the machine, and the Kokoro voice is Apache-2.0, so the video can be shared or posted.

The worked example is `${CLAUDE_SKILL_DIR}/examples/zigzag/`: 46 seconds on why gradient
descent zig-zags in a narrow valley.

## Setup

Check with `command -v manim tts ffmpeg ffprobe latex dvisvgm`. What is missing:

- ffmpeg and uv: `brew install uv ffmpeg`
- Manim Community: `uv tool install --python 3.12 manim`
- `tts`, from pdf2audio (AGPL-3.0; the voice it makes is Apache-2.0):
  `uv tool install --python 3.12 git+https://github.com/jonstraveladventures/pdf2audio`.
  Its first run downloads the Kokoro model from Hugging Face (under 400 MB).
- LaTeX, for equations (`MathTex`): MacTeX includes everything Manim uses. With BasicTeX, Manim's
  installation page lists the extra packages, installed with `sudo tlmgr install`, which asks for
  the Mac's password.

If `~/.local/bin` is not on PATH after a `uv tool install`, `uv tool update-shell` adds it. Ask
before each download and say roughly how large it is. Leave anything that asks for a password to
the user, in the Terminal app.

## Before animating: settle what it says

A polished video persuades whether or not it is right, so it only ever presents results that
were checked elsewhere, and it is never evidence for them.

- Write the claims down first, each with its source (a derivation, a paper's text, a computed
  number). Recompute every number the narration or the screen states from that source.
- A video about someone else's paper works from what the `papers` plugin's licence check allows:
  for a subscription paper, the title, abstract, short quotes and the user's own notes, never its
  full text.
- Ask what it is for if that isn't clear: a course (accuracy and pace for students), a talk, or
  the user working something out. The audience sets the length; one idea per minute is plenty.

## Pipeline

Work in a folder in the user's project, never in this skill's folder. The scripts are in
`${CLAUDE_SKILL_DIR}/scripts/`; call that `$X` below. Copy `$X/narrated_scene.py` into the
project folder first, since each scene imports it from beside itself.

1. **Narration.** Write `narration.json`: a list of `{"id", "text"}` lines in playing order, one
   line per visual beat, usually 1 to 3 sentences. Write for the ear: short sentences, plain
   words, no em-dashes, numbers as they should be said ("zero point one eight"), symbols in words
   ("eta", not "η"). Then run

       python3 $X/narrate.py <folder>

   which makes `audio/<id>.wav` and `audio/durations.json`. Rerun it after any change to the
   text. `--voice` picks another Kokoro voice (`tts --list-voices`).

2. **Scene.** Write `<name>.py` in the folder with a class that subclasses `NarratedScene`; the
   example's first lines show the import. Call `self.segment("<id>", [(animation, share), ...])`
   once per narration line: the line's audio starts there, the animations run one after another,
   each taking `share` of the line, and the rest of the line is a pause. Shares add up to 1 or
   less. Build everything the narration describes inside the segment that says it.

3. **Render** without the cache, always:

       manim -qm --disable_caching --media_dir <folder>/media <folder>/<name>.py <Class>

   `-qm` is 720p30 for drafts; `-qh` is 1080p60 for the version that goes out. The pydub
   `SyntaxWarning` lines at start-up are harmless.

4. **Check** before showing anything:

       python3 $X/check_video.py <folder> <path to the .mp4>

   It fails if the audio track is shorter than the narration, and writes `contact-sheet.png`
   with two frames per line. Read the sheet and look at every frame for text touching a curve,
   an arrow or another label, labels running off the frame, and anything shown before the
   narration names it. Fix, re-render, re-check. State in the report which frames were inspected;
   the sheet samples frames and is not the whole video, and nobody has listened to it unless
   someone did.

5. **Deliver.** Copy the final render next to its script with a plain name and tell the user
   where it is.

## Layout that avoids the usual faults

- Keep zones: title at the top edge, the main diagram in the middle, equations bottom-left,
  derived results bottom-right. Text never goes inside the plot area.
- Use `Axes` with equal x and y scale when geometry matters (contours, angles); set `x_length`
  and `y_length` in proportion to the ranges.
- Fade out a helper (an arrow, a highlight) before drawing what it would sit on.
- Dark background with theme colours set once at the top of the file; `MathTex` with explicit
  `color`, since its default is white.

## Known failure modes

- **A cached re-render drops sound.** A re-render that reused cached partial movies kept the
  full video but lost about 10 seconds of narration. `--disable_caching` prevents it, and
  `check_video.py` catches it.
- **Overlaps.** The first render of the example drew the gradient arrow on top of the first
  step's line. Code written blind gets positions wrong; only the frames show it.
- **Timing drift.** If a scene plays animations outside `segment()`, the narration and the
  picture slip apart, and the contact sheet's frame times are off by the same amount. Keep every
  animation inside a segment.
Files in this package (6)
  • SKILL.md 6 KB
  • examples/zigzag/narration.json 1 KB
  • examples/zigzag/zigzag.py 3 KB
  • scripts/check_video.py 2 KB
  • scripts/narrate.py 2 KB
  • scripts/narrated_scene.py 2 KB

Versions

RepositoryScoreLicence
This page · shocklab/co-scientist-toolkit100MIT
smykla-skalski/sai100MIT
fengchen122123-sudo/xingyao-y-code100MIT
menschenkreis/explainer-video100MIT
0xsline/OpenChatCut100AGPL-3.0
Prahaladha-Reddy/skills100MIT
Andyyesiyu/explainer-video-skill100MIT
mys1erious/skills100MIT
baskb/explainer-video100MIT

Database snapshot: 13.11.2026 · package version 8b10646c073c (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100manim
    Make a narrated explainer video with Manim (3Blue1Brown style): script it, animate it, add an ElevenLabs voice and phrase-by-phrase captions, check the layout, render 1080p, and deliver the MP4. Use when the user asks for an explainer/animated video, says "make a video about X",…
  • 100manim-explainer
    Create short math and science explainer animations with Python Manim: from a concept to a scene plan, a minimal project, a render command, and a verified output file. Use when the user wants a Manim animation, a math explainer video, or says "animasi manim", "buatkan animasi…
  • 100manim-video
    Build reusable Manim explainers for technical concepts, graphs, system diagrams, and product walkthroughs, then hand off to the wider ECC video stack if needed. Use when the user wants a clean animated explainer rather than a generic talking-head script.
  • 100manim
    This skill should be used when the user asks to "make a math animation", "animate an equation", "use Manim", "make a 3Blue1Brown-style video", "plot or animate a graph/function", "visualize a math/physics/CS concept", "morph one formula into another", "animate a geometric…
  • 95community-manim-video
    Production pipeline for mathematical and technical animations using Manim Community Edition. Creates 3Blue1Brown-style explainer videos, algorithm visualizations, equation derivations, architecture diagrams, and data stories. Use when users request: animated explanations, math…
  • 84manim-video
    Production pipeline for mathematical and technical animations using Manim Community Edition. Creates 3Blue1Brown-style explainer videos, algorithm visualizations, equation derivations, architecture diagrams, and data stories. Use when users request: animated explanations, math…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.