Trialled by Daniel Zuboff
Strategy at Loomery
Adding videos to a website or a blog post is a great way to make your content stand out. People spend 88% more time on websites with videos vs those that are just text-based. Unfortunately, creating video content is often a laborious process, involving scripting, recording a voice-over, finding videos or images to capture someone’s attention, and editing the whole thing. Fliki is an AI powered tool that makes creating such content incredibly easy.
It generates a realistic sounding voice, it can scan the text you input to automatically select appropriate video clips that mirror the points made in your script, and export a video file complete with subtitles and music. While these videos are formulaic and predictable, a lot of the time, I’d prefer to get my information delivered this way, especially when used in combination with LLM’s like GPT to accurately summarise larger chunks of text.
For a brief demo, I fed Loomery co-founder Tim’s blog post “The loom: programming patterns on a path to computing” into ChatGPT, asking it to generate a short script; a job it performed very well.
I then pasted this text into Fliki, picked a style of voice along with semi-appropriate clips and images and the video was ready to be downloaded.
The AI didn’t do a good job of selecting appropriate clips, or pronouncing certain words; it repeatedly got “Jacquard” wrong. The library to choose clips and images from was also very limited but it did save a lot of time vs searching for relevant clips, downloading them and editing them together. With the paid version of the service, you can correct mispronunciations and browse through more clips but in general you do get out what you put in. This kind of low-effort content creation is pretty transparent now, but I’m sure these kinds of tools will only improve over time.
https://drive.google.com/file/d/1jTL4CuAvurHH7ZmU9lczfdQvP16OHC0T/view?usp=share_link