back to labs
lab experiment LAB-004 content automation · 2025 remotion · mcp · canva

programmatic video
& automation.

Code-driven video templates with Remotion and automated Canva carousel generation through Claude Code and MCP. From template to published content — no manual design work.

see it in action

the walkthrough

A complete walkthrough of both automation workflows — Remotion video template generation and the Claude Code + MCP Canva carousel pipeline — from structured content input to published output.

the experiment

two workflows. one automation blueprint.

This experiment combines two separate automation workflows — each solving a different layer of the content production problem.

01

remotion video templates

Remotion turns video production into a code problem — version-controlled, parameterized templates that generate consistent branded video without touching a timeline editor. Change the copy, re-render, done.

02

canva mcp carousel generation

The Canva MCP integration takes a content queue from Notion, reads designated templates, and builds every carousel automatically. Monday morning, the carousels exist — nothing was touched manually.

for L&D teams

This is the blueprint for scalable asset production — job aids, micro-videos, onboarding materials, and visual guides all generated programmatically from structured content, not manual design work. The same logic that builds a carousel builds a job aid template.

the output

7 slides. zero manual design.

hover any slide to pull it forward

carousel slide 4
carousel slide 3
carousel slide 2
carousel cover slide
carousel slide 5
carousel slide 6
carousel slide 7
tools used

the stack

remotion claude code mcp canva notion react

limitations & what's next

Early-stage automation with clear expansion paths.

template variety

Currently handles a focused set of carousel and video templates. Expanding to story templates, infographics, and short-form video variations.

multi-platform output

Video templates currently target a single aspect ratio. Next: automatic rendering for vertical (Reels/TikTok), square (feed), and widescreen (YouTube) from one template.

ai-driven adaptation

Templates use structured content input. Next: AI that rewrites and adapts copy for different audiences and platforms automatically before rendering.

review & approval gates

Currently outputs go straight to done. Adding a preview and approval step so generated assets get reviewed before they're marked ready to publish.

next in the lab

context-aware
microlearning.

An interactive demo that simulates moment-of-need learning delivery — the right nudge at the right moment, not just shorter content.

try the experiment