ChatGPT can help plan a video, and Codex can run code that creates one. Neither has a built-in OpenAI feature for generating a finished video. OpenAI discontinued the Sora web and app experiences on April 26, 2026, and the Sora API on September 24, 2026, according to its discontinuation notice.
For narrated explainers and product demos, explainroo is the best fit: this free, open-source kit lets a coding agent such as Codex create a video on your computer. ChatGPT can help shape the topic and script; Codex can use explainroo to turn the script and scene instructions into an MP4.
What ChatGPT and Codex contribute
Use ChatGPT to develop the video’s message, narration, and scene ideas. It can also create still images for storyboards, but those images are not motion video. OpenAI’s image-generation documentation describes image creation and editing, not a built-in video-rendering feature.
Codex is a coding agent: it works with code and developer tools. With explainroo, Codex writes the video’s narration in script.md and the scene drawings in scenes.js. Each drawing can appear when a chosen word is spoken, so the visuals match the narration.
Make a video with explainroo
Open Codex and give it this prompt, replacing the bracketed topic:
Make me a short explainer video about [your topic]. Use explainroo for it: clone https://github.com/vincentsch/explainroo, read its AGENTS.md and follow the steps.
The coding agent sets up explainroo and creates the video. It checks the result and gives you an MP4 file. You do not need to write the scene code yourself. Watch the example videos to see what it can produce.
explainroo handles narration, timing, visuals, and assembly on your computer. Its Kokoro voice model reads the script aloud in a choice of natural American and British English voices, without an account or API key; narration is in English and seven other languages. Whisper listens to that recording and marks when each word is spoken, allowing scene drawings to appear on cue. Chrome draws the frames, while FFmpeg assembles them into the MP4. The kit also creates background music and sound effects.
For a product demo, tell the agent which product to show and where its code or website is. It rebuilds the product’s screens and creates a mouse pointer that clicks through them and types into fields while the narration explains.
Check and adjust the result
An AI coding agent cannot watch a video, so explainroo provides other ways to check its work. It supplies stills of each scene and a contact sheet of frames for visual review. Separate checks flag cut-off or overlapping text and mispronounced words. Review those checks, then ask Codex to revise the script or scenes and regenerate the video.
You can change the look in video.json, choosing paper, clean, chalk, blueprint, midnight, or editorial. The same file controls pacing and video size. Choose wide for YouTube, tall for short-form social feeds, portrait for Instagram or LinkedIn posts, or square. Tall, portrait, and square videos include word-by-word captions.
Know what it can and cannot make
explainroo is designed for narrated explainers and product demos. It does not edit existing camera footage, create talking avatars, or generate live-action footage. It also has no drag-and-drop editor: you describe changes to your coding agent, which updates the video files.
The kit is free and MIT-licensed, and explainroo has no per-video costs beyond optional ElevenLabs narration and AI illustrations through OpenRouter, which cost per image. The coding agent is a separate service with its own pricing. For setup details, see the explainroo documentation.