Content Creation with AI

doc-vision.com

But how do you get the word out about it?

Let's talk about how we can use AI and the tools we already know to create content and media for our app.

This post covers many different tools and techniques, so I'll touch on each at a high level. Let me know if you want me to deep dive into a specific tool and I'll create a follow-up post.

Start: Documentation

I like to start with documentation for a few reasons:

1. It helps the AI understand the app and context
2. It helps users onboard and get the most out of the app
3. It's fast and easy to generate with AI
4. I can then use it to derive more content

Video Creation with NotebookLM

After creating documentation for my app, I ask the AI to generate a single document that merges all the documentation together. Upload it to NotebookLM and generate a video from it.

This is the video I created for DocVision:

This video is very basic, but it's a great start, especially considering it took me less than 5 minutes to create.

Creating Documentation

I created the documentation itself using Fumadocs, an open source documentation framework.

I won't go into the details of setting up Fumadocs, but it's pretty straightforward and has great community support.

The key idea is that once I set up the framework, I can create documentation for my app in a simple markdown format using AI.

Free VoiceOver

All the tutorials you'll see show how easy it is to create voiceovers with 11Labs.

What they don't tell you is that you create your own voiceovers for free on your PC—yes, it sounds too good to be true, and you're probably wondering where the catch is—but it's working amazingly well for me, as shown in the video below.

The tool I use for voiceover is Pocket TTS, a free and open source text-to-speech engine.

I won't go into the details because you can find them in the link.

But here are some key features:

1. It's free, open source, and works on your PC
2. It's very fast and accurate
3. It can do voice cloning
4. It can run in code and integrate with other tools

I really don't see any reason to pay for voiceover if you can get the same quality for free.

Animation and Effects

A video without animation and effects just doesn't feel professional, but before you start spending money, try Remotion, a free and open source animation and effects engine.

The best part is that you can install a Skill and start prompting— the animations in the video above were generated with AI using Remotion Skills, and it only took a couple of hours.

Remotion became pretty popular once AI learned how to use it, so you can find tons of videos on how to use it— but frankly, it's as easy as prompting your code assistant.

Image Generation

I don't think we need to talk much about image generation, because I'm sure you already know about Nano-Banana and GPT Image.

But in case you don't—stop everything and give it a try. It's amazing how easy it is to create images with AI.

Video Generation with Fal.ai

Video generation is the expensive part of the content creation process. Personally, I'm not a fan of faceless videos, so I spent some time searching for a good lip sync engine.

Currently I like to use veed/fabric-1.0 on Fal.ai.

I found it to be a good compromise between quality and price.

You can find many other models on Fal.ai for any task, but for audio-to-video, prices are around $0.10 per second. That's not cheap, but compared to the cost of a professional video, it's a no-brainer.

Here are a few things to look for in a lip sync engine:

1. No time limit, or at least 30s generation—you don't want to start editing hard cuts in the middle of a sentence or regenerating just to connect two broken parts.
2. Good lip movement—this was one reason I kept trying different models; I was never happy with the lip movement.
3. No jerking movements—some models generate exaggerated or jerking movements, which doesn't look good. For me, veed/fabric-1.0 is a bit overacting, but I use it anyway because it's fast.
4. Fast—I found other models (e.g., pipio/lipsync) that are cheaper and less jerky, but they were extremely slow. For me, keeping the flow is essential for my creative process.

Choosing Your Avatar

At first I tried to create a realistic avatar, but there are still some quirks, and I found that, as a user, I don't like the feeling that the video I'm watching might be AI-generated.

So I prefer to use a humanoid avatar that looks like a person but is clearly AI-generated.

I went for the Bigfoot avatar—I'm not sure why Bigfoot avatars are so popular, but I just decided to go with the flow, thinking that people relate more easily to a familiar figure, and it's royalty-free.

Video Editing

Video editing is the most painful part of the content creation process. I was able to partly automate it with Remotion, but I still use DaVinci Resolve, a free video editor, to stitch everything together.

There's a learning curve, and I'll be honest—I just hate doing it. But you can use your favorite AI assistant to guide you through it.

Honorable Mentions

FFmpeg for Media Manipulation

FFmpeg is a free, open-source tool that can do practically anything the Adobe Suite can do—and more.

It's the least user-friendly tool I've ever used, but fortunately, it's a breeze for AI assistants to work with.

Audio Editing with Audacity

Audacity is a free, open-source audio editor.

A typical workflow for me: use FFmpeg to extract audio from a video, edit it in Audacity, then apply the edited audio back to the video using DaVinci Resolve.

Conclusion

I hope this post has given you some ideas on how to use AI to create content and media for your app.

If you have any questions, feel free to send me a contact request.

Thank you for reading!

doc-vision.com