<aside> 💡 Install the Extension: While our web app supports PDF uploads, we also offer a Chrome extension that lets you generate videos from any webpage with just one click.

</aside>

Supported videos 🎞️

The current version of our video AI technology works best for the following types of content:

• Educational courses
• How-to guides and tutorials
• Articles and blog posts (e.g. Substack)
• Academic papers (not too technical)
• Case studies with insights (not too analytical)
• Informational content (e.g. Wikipedia, Britannica)

<aside> ❗ We haven’t prioritised content pieces that lack any form of educational value, such as pure entertainment, lifestyle content, or celebrity pieces. However, you can still generate videos for them if you wish.

</aside>

<aside> 🔥 We are actively working on supporting technical and analytical content. You can already generate coding tutorials, but we still don’t support math or complex data visualisations for financial articles or other related use-cases.

</aside>

Best practices 🙏

We’ve put together some best practices to help you improve the quality of your video output:

<aside> 📹

Use Lumina v3.5: For most use cases, we recommend using our latest model, Lumina v3.5 – it’s our best and most advanced model and performs better for almost all areas of knowledge.

</aside>

• Try different voices: Videos improve significantly with our premium, human-quality voices. However, not all voices are equal — some perform better than others depending on the content type, target audience, and terminology used.
• Submit smaller pieces whenever possible: For most educational use cases, bite-sized videos work better for learning and result in higher-quality outputs.
• Include useful images, diagrams, and code snippets: If your content contains high-quality visuals or code snippets, keep them in the right places. The system will try to use them effectively.
• Avoid unnecessary text, images, or icons: Extra illustrations or text that don’t add value can confuse the model and lower the output quality. Keep your content as clean as possible. If your PDF comes from a presentation or design-heavy guide, consider copy-pasting the text and only the important visuals into a separate document before generating the video.
• [NEW] Custom Instructions: You can tell the AI exactly how long you’d like your video to be, how you want specific parts of your content processed and how to structure the videos. Note: This feature is available with Lumina v3.5.

How it works Just add instructions anywhere in your document using this simple format: [INSTRUCTION: your instruction here] That's it. The AI will follow your custom instructions across the document. You can add as many instructions as you need, wherever you need them. Just make sure you strictly follow this format.

**What you can control**

Here are some practical examples:

- **Video length:** `[INSTRUCTION: Make the video [X] minutes long]`
- **Emphasize sections:** `[INSTRUCTION: Focus heavily on the benefits section]`
- **Structure the flow:** `[INSTRUCTION: Start with an introduction about customer pain points]`
- **Highlight quotes:** `[INSTRUCTION: Include this as a text snippet: "Quality over quantity"]`
- **Condense content:** `[INSTRUCTION: Summarize the technical details into one scene]`
- **Expand topics:** `[INSTRUCTION: Break down the implementation steps into multiple scenes]`
- **Control naming:** `[INSTRUCTION: Refer to this as "AI Agent" not "Assistant"]`
- **Explain details:** `[INSTRUCTION: Spend enough time explaining all the details in the following table]`
- **Guide image use:** `[INSTRUCTION: Use the image above as an example of the final result—focus on what it demonstrates]`
- **Highlight names:** `[INSTRUCTION: Make sure to highlight Dr. Sarah Chen as the lead researcher]`