Persian Motion Director logo

Persian Motion Director

Community
atmirrr
persian-motion-director

Direct and build professional motion design (promo and launch films, kinetic typography, animated explainers, social clips, title sequences) with correct Persian/Farsi and right-to-left typography, in Remotion or plain web code. Use this skill whenever the user wants a video, animation, motion graphics, teaser or animated text that contains Persian, Farsi, Arabic-script or RTL text, and whenever they say a motion piece looks "default", "like a slideshow", "AI-made" or "not pro", or that the Persian looks off, broken or disconnected, even if they never say "motion design". It brings the anti-slideshow method (one camera, objects that carry across scenes, music-first beat-locked edits), eight ready techniques with live sketches, HarfBuzz tooling for per-letter and per-dot Persian animation, kashida stretching, and a Persian correctness checklist and linter.

Overview

Publisheratmirrr
Repositorypersian-motion-director
Skill namepersian-motion-director
Stars
147
Forks
6
Bundled files
56
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 56 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by atmirrr on GitHub. Read the source before you install it.

Installation

Install the Persian Motion Director AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/atmirrr/persian-motion-director/tree/main
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/atmirrr/persian-motion-director.git \
  .claude/skills/persian-motion-director
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Persian Motion Director in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Persian Motion Director on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Persian Motion Director is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Persian Motion Director

You are directing, not just animating. Most generated motion fails in two predictable ways: it plays like a slideshow (scenes that fade into each other, one ease-in-out on everything, everything arriving at once), and its Persian is broken (letters split apart, dots clipped, Arabic letters and digits, missing half-spaces, left-to-right reveals). This skill fixes both. The method comes from how professional motion designers describe their own work on X; the Persian rules come from W3C's Arabic & Persian Layout Requirements and were checked in Chrome (the engine Remotion renders with).

The director's loop

Work in this order. Each step exists because skipping it is how pieces end up looking default.

  1. Read the brief and fill only real gaps. Copy (Persian and any Latin), length, aspect ratio, brand colours and fonts, where it will be posted, music. If a brand exists, use its palette and fonts instead of inventing one.
  2. Show directions before building. Motion is subjective and expensive to redo. Pick 6–10 techniques from the library below that fit the brief, open assets/sketches/gallery.html (or render the clips) and let the user choose. Rework after a full build costs ten times more than a pick list.
  3. Music first. Lock the track and its tempo before animating. Build a bar map: which section lives on which bars. If there's no track, compose one (assets/music/compose_shur_pulse.py is a working example that writes beats.json) or ask for one. At 120 BPM one beat is 0.5 s and one bar 2 s; at 60 fps a beat is 30 frames.
  4. Write the shot list on the bar grid (template in references/workflow.md). Every section change lands on a downbeat; every cut on a beat or 2 frames before it.
  5. Style frames. Render 4–6 stills (assets/remotion/render-stills.mjs) and get a yes before animating everything.
  6. Build in Remotion with one camera container and shared objects (see assets/remotion/src/intro/Intro.tsx for a complete worked example).
  7. Check, then render: run scripts/lint_persian.py on all copy, review stills at full size for clipped dots and broken joins, scrub every cut against the beat, and make loops exact (clip length = a whole number of loops).

What separates pro motion from a slideshow

Slideshow habitWhat to do instead
Scenes chained with fade transitionsOne camera travels between scenes, or one object carries into the next scene. Never fade() scene to scene.
The same ease-in-out on everythingEach move gets its own curve. Entrances: bezier(.16,1,.3,1) over ~20 frames. Exits: bezier(.7,0,.84,0) over ~8 frames, travelling only ~70% as far.
Everything arrives together, evenly spacedOne lead element, the rest 2–4 frames behind, spaced on a curve rather than evenly.
Remotion's default bouncy spring() (overshoots ~16%)Under 2% overshoot on UI, none on text: stiffness 100 / damping 16 gives ~1.5%.
Music laid on at the endMusic first; cuts on the beat; sound attacks placed on the visual hits.
Centered text on a gradient, logo at the endShow the real product doing a real task in the first 10 seconds.
Freezes, or motion that never pausesHold long enough to read; only the camera breathes (zoom 1.00 → 1.04 per scene).
Grain, halftone or ASCII as a finishTexture only with a physical reason (paper, print, light). These finishes are now the AI default.

More numbers and the sources behind them: references/anti-slideshow.md.

Persian non-negotiables

Read references/persian-typography.md before writing any Persian into a frame. The short version:

  1. Never put Persian letters in their own boxes. inline-block spans, flex children (including direct children of Remotion's <AbsoluteFill>, which is a flex box) and any per-letter transform draw each letter in its isolated form. Animate whole words or lines. When you truly need per-letter or per-dot motion, use shaped outlines (scripts/shape_persian.py) or zero-width-joiner splitting (scripts/split_fa.js).
  2. Fade words, not letters. Joined letters overlap; per-letter opacity or see-through text colour leaves seams at every join.
  3. Masks need slack. Persian reaches far above and below Latin: pad clip masks by about 0.5em top and bottom, use line-height ≥ 1.6 (≥ 2.4 for Nastaliq), or the dots and the top of گ get cut.
  4. Direction. dir="rtl" lang="fa" on every Persian block; wrap Latin names and numbers in <bdi> (or FSI/PDI) or the line scrambles. Reveals and staggers start on the right, camera progress travels left, "next" arrows point left. Play buttons, clocks and numbers are never mirrored.
  5. Characters. Persian ی (U+06CC) and ک (U+06A9), never Arabic ي/ك; Persian digits ۰–۹ (U+06F0–06F9), never Arabic-Indic ٠–٩; the half-space (ZWNJ, U+200C) in می‌, ‑ها, ‑تر/‑ترین; punctuation ، ؛ ؟ « ». scripts/lint_persian.py catches these.
  6. No letter-spacing. Chrome ignores it inside Persian words (since version 137) and other tools break the joins. Stretch with kashida instead, only after a letter that joins forward, at most one per word.
  7. Load the font, every time. If the Persian font (or its Arabic-script subset) isn't loaded, Chrome silently falls back to a system font. In Remotion use FontFace + delayRender (see assets/remotion/src/fonts.ts).
  8. Copy. Write natural Persian in one register; no word-for-word English idioms. A native reader signs off on every line before render (references/persian-copy.md).

Technique library

Live code for each is in assets/sketches/sketches.js (plain JS, a pure function of time) and runs both in a browser page and in Remotion through SketchPlayer. Details, numbers and the X posts each comes from are in references/techniques/.

#TechniqueUse it forSketch id
01One camera, no cutsThe spine of a film: scenes on one strip, whip pans with motion blur and parallaxoner
02One shape carries the storyTransitions: a full stop grows into a chat bubble, then into the next full stopmatch
03The real UI, directedProduct beats: real interface in 3D, cursor on arcs, focus pullsui
04Cut on the beatFeature runs and kinetic type: hard cuts on every beat, punch-ins, inverted framesbeat
05On twos, with line boilA hand-drawn layer over clean UI: drawings hold 2 frames, lines wobble 12×/sboil
06Paper cut-out lettersWarm title cards: each joined letter group is a rigid paper piececutout
07Dots lastWordmarks and reveals: letter bodies first, then the dots drop innuqta
08Kashida, not letter-spacingEmphasis on a held note: stretch the joining stroke, measured in calligraphic dotskashida

Pick by the brief, not by novelty: one or two signature techniques used consistently read as a system; all eight at once read as a reel.

Tools in this skill

  • scripts/shape_persian.py — shapes Persian with HarfBuzz and exports SVG outlines split into letter bodies, dots and connected letter groups (pip install uharfbuzz fonttools). This is what makes per-letter, per-dot and kashida animation possible without breaking the script.
  • scripts/split_fa.js — splits a Persian word into letters that keep their joined forms (zero-width joiners), for when outlines are overkill.
  • scripts/lint_persian.py — checks copy (Arabic letters and digits, missing half-spaces, English punctuation, misplaced kashida) and code (letter-spacing, per-letter boxes, missing dir).
  • assets/sketches/ — the motion engine (engine.js), the eight sketches, shaped glyph data, styles and gallery.html (serve the folder with any static server).
  • assets/remotion/ — SketchPlayer (drives a sketch from the frame number), one clip composition per technique, the full intro film as a worked example, and scripts to render clips and stills.
  • assets/music/compose_shur_pulse.py — an original 120 BPM track in Dastgah Shur written in code, with a beat map; shows how to make the edit lock to the music.

When the user only wants a quick piece

Still keep the two non-negotiables that cost nothing: no per-letter boxes for Persian, and no scene-to-scene fades. Use one technique well, set the timing on a beat grid, and render stills to check the Persian before the full render.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Persian Motion Director AI skill do?

Direct and build professional motion design (promo and launch films, kinetic typography, animated explainers, social clips, title sequences) with correct Persian/Farsi and right-to-left typography, in Remotion or plain web code. Use this skill whenever the user wants a video, animation, motion graphics, teaser or animated text that contains Persian, Farsi, Arabic-script or RTL text, and whenever they say a motion piece looks "default", "like a slideshow", "AI-made" or "not pro", or that the Persian looks off, broken or disconnected, even if they never say "motion design". It brings the anti...

Why use Persian Motion Director on TypingMind?

Because you install it once and use it with any model. Persian Motion Director is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Persian Motion Director in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/atmirrr/persian-motion-director/tree/main. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Persian Motion Director?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Persian Motion Director?

As many as you like. As long as a model supports skills, you can use Persian Motion Director with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Persian Motion Director AI skill free?

It is published on GitHub by atmirrr. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇