Explainroo is a video generation tool that turns scripts into narrated explainer videos and product demos by combining local text-to-speech, word timing analysis, and canvas rendering to produce MP4 files.
The tool solves the problem of creating explainer videos without manual animation or video editing. A coding agent writes two files: a markdown script containing narration and a JavaScript file that draws scenes. Explainroo then synthesizes speech using Kokoro, an open-source voice model with 28 voices that requires no API key. Whisper analyzes the audio to determine precise word timing. Chrome renders the scenes in the background, supporting hand-drawn line styles through Rough.js, access to 1,800 Lucide icons, charts, code snippets, and custom screenshots. The tool generates background music that automatically ducks during speech, adds sound effects, and uses ffmpeg to assemble everything into a final video. All processing happens locally on your computer at no cost.
Explainroo is designed for developers who want to automate video creation through coding agents like Claude Code, Codex, or Pi. It suits projects where an AI agent can generate both the script and scene code, then verify the output through layout checks and speech validation. The tool provides visual feedback through still frames and a contact sheet of the entire video, plus checks for text overflow and speech recognition errors. It offers five visual styles—paper, clean, chalk, blueprint, and midnight—selectable through a single configuration setting. Optional AI-generated illustrations are available through OpenRouter at per-image cost.
The project maintains active engagement with coding agents as its primary interface, with documentation specifically addressing agent integration. Development shows responsiveness to agent capabilities, with notes on which models work best. The tool includes comprehensive checking and validation features that acknowledge the constraint that agents cannot watch videos directly. The codebase integrates multiple specialized components—Kokoro, Whisper, Chrome rendering, and ffmpeg—into a cohesive workflow designed around agent-driven workflows.
