ASCII Art Video

About ASCII Video

Terminal-grade motion graphics, the character-grid way: a monospace character grid in three tones — dim scaffolding, cream body text, coral accents that glow. No video, no images — every frame is drawn from characters alone, and the picture stays exactly in step with the timeline whether you're scrubbing, editing or exporting.

The look lands somewhere between a boot log and a broadcast title card, and it suits exactly the kind of video where a diagram would be too literal and stock footage would be a lie: explaining how a model samples the next token, what a context window actually holds, why an industry piles onto its own.

What your AI agent can build here

The template seeds the ascii-terminal skill, so any AI agent working in a project made from it already knows the format and its rules:

• Boot sequences and typewriter lines — cursor-tracked typing at a set characters-per-second, with the caret behaving the way a real terminal's does.
• Scramble-reveal titles — a 5-row block font whose letters resolve out of noise, one staggered glyph at a time.
• Self-drawing panels — boxes whose borders draw themselves clockwise, rails and header/footer chrome, tick marks and gauges.
• Live data figures — probability bars, dials that reshape live odds, scrolling chat history, progress meters and density fields built from shaded characters.
• Character wave fields and sprites — multi-row ASCII art, drifting fields, and scatter that stays put — the same moment always draws the same frame, take after take.
• Narration sync — the standard voice loop: one paragraph per beat, TTS parts rendered by the audio plugin and placed on the timeline, then scenes retimed to the real voice durations.

How to work with it

1. Use this template and give the project a name.
2. In the project window, describe the video — "a 30-second explainer on context windows: chat history scrolling off the top, then a compaction event".
3. Scrub the stage while your agent works; the preview catches up within a beat.
4. Add narration once the scenes settle, then export.

The bundled samples are finished, narrated films built exactly this way — open one as a project to see complete scene functions, a voice manifest, and a timeline wired together.