Problem / motivation
We'd love OpenMontage to be a great tool for producing educational explainer videos — the kind of clean, infographic-rich, motion-graphics content you see from Kurzgesagt, TED-Ed, and The Infographics Show, adapted for programming, AI, and DevOps topics. Think animated code walkthroughs, architecture diagrams that build themselves, AI chat demos, and data-driven infographics, running up to about 10 minutes.
The foundations are already here. A handful of additions would take it from "possible with workarounds" to "excellent out of the box."
Motivation
Technical education is a huge and growing use case, and OpenMontage is unusually well-positioned for it: it already has an explainer pipeline, animated chart components, and genuine pedagogy theory baked into its skills. Today, though, two things get in the way:
- Long-form is hard to reach. The system is comfortable at 30–180 seconds. A 10-minute educational video is technically possible but pushes against short-form defaults and limits.
- The signature visuals are static. Code and diagrams render as still images. To feel like a modern explainer, they need to move — typed code, diagrams that reveal step by step, animated infographics.
Closing these two gaps would unlock a whole category of high-value content.
What already works (great starting point)
- Long-form playbook —
skills/creative/long-form.md covers 8–15 min structure: chapters, pacing, retention, audio targets.
- Animated data visualization — real animated bar / line / pie / KPI / stat / comparison scenes, plus a strong
data-visualization skill. This is the best-supported area today.
- Pedagogy grounded in research —
skills/creative/storytelling.md draws on Mayer's Multimedia Learning, misconception-first teaching, and guided-discovery structure.
- Synthetic terminal & UI demos — deterministic, privacy-safe CLI and screen demos, ideal for DevOps content.
- Concept animation — Manim support for AI/ML math (gradients, vectors, embeddings).
Proposed solution
1. Make up-to-10-minute videos a first-class option
- Offer a long-form option at the idea/proposal stage (e.g. 8–15 min, chaptered) and route it to the existing
long-form playbook, which nothing currently references.
- Use duration-aware defaults for budget, wall-time, and render timeouts so a 10-minute run doesn't hit short-form ceilings.
- Handle long music beds reliably (generate up to the provider limit, then loop/crossfade to full length).
- Add a safety net for long narration so the pipeline degrades gracefully instead of failing on length limits.
2. Turn static assets into animated motion-graphics scenes
- Animated code walkthroughs — an editor-style scene with syntax highlighting, typed reveals, and line-by-line focus (not a static screenshot). This is the single most impactful visual for programming content.
- AI chat / agent transcript scene — chat bubbles with typing animation, for showing prompts and model/agent behavior in AI explainers.
- Animated diagrams — architecture and system diagrams that reveal nodes and connections step by step, rather than appearing all at once.
- Composite infographic scene — a designed board that layers a number, an icon, a label, and a mini-chart into one animated panel — the classic "infographic" look.
- General annotation layer — arrows, circles, and highlight boxes that can be drawn over any code, diagram, or chart to direct attention.
3. Teaching structure and polish
- A pedagogy skill that shapes technical explainers around learning: state the objective, show a worked example, call out the common pitfall, recap. Introduce one new concept at a time.
- A PR / diff review scene for code-review and DevOps walkthroughs.
- Kinetic typography as an option for emphasis text within the main runtime.
Alternatives considered
No response
Additional context
- Reference style: Kurzgesagt, TED-Ed, The Infographics Show — clean, modern, motion-graphics-driven explanation.
- Terminology: infographic = the static visual; motion graphics = the animated version. The heart of this request is moving technical visuals from static infographics toward animated motion graphics.
Problem / motivation
We'd love OpenMontage to be a great tool for producing educational explainer videos — the kind of clean, infographic-rich, motion-graphics content you see from Kurzgesagt, TED-Ed, and The Infographics Show, adapted for programming, AI, and DevOps topics. Think animated code walkthroughs, architecture diagrams that build themselves, AI chat demos, and data-driven infographics, running up to about 10 minutes.
The foundations are already here. A handful of additions would take it from "possible with workarounds" to "excellent out of the box."
Motivation
Technical education is a huge and growing use case, and OpenMontage is unusually well-positioned for it: it already has an explainer pipeline, animated chart components, and genuine pedagogy theory baked into its skills. Today, though, two things get in the way:
Closing these two gaps would unlock a whole category of high-value content.
What already works (great starting point)
skills/creative/long-form.mdcovers 8–15 min structure: chapters, pacing, retention, audio targets.data-visualizationskill. This is the best-supported area today.skills/creative/storytelling.mddraws on Mayer's Multimedia Learning, misconception-first teaching, and guided-discovery structure.Proposed solution
1. Make up-to-10-minute videos a first-class option
long-formplaybook, which nothing currently references.2. Turn static assets into animated motion-graphics scenes
3. Teaching structure and polish
Alternatives considered
No response
Additional context