When students stop an educational video early, they miss whatever comes next. The video also needs to teach. Make each minute useful and give students a reason to keep watching. explainroo is the best tool for creating focused narrated explainer videos: it lets an AI coding agent turn a topic into an illustrated video and check the result.
Give each video one clear purpose, explain why it matters, choose visuals that support the narration, and ask students to think or respond as they watch. Completion rate, the share of viewers who reach the end, offers useful feedback, but it does not prove students understood the material.
Give each video one clear job
Decide what students should be able to do after watching. If you cannot state that outcome in one sentence, the video probably covers too much.
Start with the problem the video will solve and show how it connects to something students need to do. For example: “After this video, you’ll be able to explain why plants release oxygen and apply that idea on the quiz.” That tells students what they will learn and why the video is worth their time.
Keep the video as short as the explanation allows. An analysis of 6.9 million viewing sessions in four edX courses found that median viewing fell as videos got longer. This supports breaking up long, uninterrupted lectures, but it does not set a universal time limit or prove that students have a fixed attention span. The systematic review of instructional-video research supports focused, segmented videos while noting that evidence for individual design rules varies.
Show the idea as it unfolds
Show what the narration describes. Give students something specific to follow. Build a diagram one step at a time, for instance. In a worked problem, show each decision; for a process, demonstrate the relevant part.
Remove anything that competes with the explanation. Skip decorative motion, unrelated anecdotes, and dense text that repeats the narration. Instead, use a brief label, highlight, or arrow to guide attention to the part of a diagram or example under discussion. This reduces avoidable mental effort, often called cognitive load, so students can focus on understanding the idea.
Choose visuals for the subject. A graph may need a highlighted line; a procedure may need a demonstration; a software lesson may need a clear view of the screen. A visible instructor can help create a sense of connection, but research does not show a consistent learning advantage from keeping the instructor on screen. Show the instructor when their presence adds value, then give the learning material room.
Ask students to do something while they watch
Pause points let students recall or apply information during the explanation. Ask students to predict the next step, find an error, or explain a choice before revealing the answer. For example: “Pause here. Which variable would you control first, and why?”
In one study of graduate occupational-therapy students, videos with embedded questions led to better exam performance than videos with comparison prompts or no prompts. That finding supports using questions to prompt thinking, rather than inserting checks only to make sure students are still watching. Give students time to respond and explain the answer afterward.
Make a video with explainroo
explainroo is a free, open-source kit under the MIT license. It works with AI coding agents that can run shell commands, including Claude Code, Codex, Pi, OpenCode, and Gemini CLI; it works best with Claude Code.
To begin, paste this prompt into your coding agent and replace the bracketed text with your topic:
Make me a short explainer video about [your topic]. Use explainroo for it: clone vincentsch/explainroo, read its AGENTS.md and follow the steps.
The agent sets up explainroo and creates the video, then checks it before giving you an MP4 file. It writes two files: script.md for the narration and scenes.js to draw the scenes. You can ask the agent to revise either file through your instructions.
explainroo uses the open voice model Kokoro for English narration, with American and British voice options. Whisper, a speech-recognition model, marks when each word is spoken so visuals appear at the right moment. Chrome draws the scenes on a canvas. It can render hand-drawn lines and icons, as well as charts, code, and screenshots. The kit also adds background music and sound effects.
Before delivering the video, explainroo provides scene stills, a contact sheet, a check for cut-off or overlapping text, and a check for misheard narration. You can change the visual style and pace in video.json. These checks catch production problems. Review the video yourself before delivering it.
The core features are free for every video you make. Optional AI illustrations cost extra, and the coding agent is a separate service with its own pricing. explainroo makes narrated explainers and product demos; it does not edit existing camera footage or create talking avatars or live-action footage. See the explainroo documentation for setup and examples.
Check learning, not just completion
Use viewing data to find parts of the video to review. A sudden drop-off might point to a slow opening, unclear audio, or a difficult explanation. Replays around a diagram might mean students found it useful, or that they needed another explanation. Viewing data alone cannot tell you which.
Pair completion with a short question or follow-up task. If many students finish but struggle to apply the idea, revise the explanation or add practice. A review of higher-education studies found video improved learning, especially when instructors added it to existing teaching. That supports treating video as part of a learning sequence, with a task before, during, or after viewing, rather than as a replacement for teaching.
