Videos are processed one-by-one by our video rendering servers and they work on a first-come-first-serve basis. For example, if 5 people submitted their videos for rendering before you did, then our servers must finish processing those 5 videos first before it can start processing your video. As you can imagine, this can easily create a bottleneck when a lot of users are on Lumen5 creating videos at the same time.
Previously, we were running a single rendering server, which was why your videos could sometimes take over an hour to render. To resolve this, we decided to rebuild our backend infrastructure from ground up. Now, we are able to run multiple rendering servers in parallel with the ability to easily expand our rendering capabilities on-demand to handle traffic surges.
With the new setup, we can now render multiple videos in parallel. In our earlier example, if 5 people submitted their videos before you submitted yours, those 5 videos must be completed first before the server begins processing your video. Now, all 6 videos can be rendered at the same time, and everyone can get their videos sooner.
The average rendering time is now consistently under 10 minutes and we plan to continue improving our system in order to process videos even faster in the future!
You can now share your Lumen5 videos directly to Facebook! Rather than downloading the video file to your computer and then uploading it to Facebook manually, you can now use our Facebook Upload integration to share the video directly to your page or timeline.
When you use the integration for the first time, Facebook will ask you for permission for Lumen5 to post on your behalf, so please make sure you Agree! The integration will only share videos when you click the Upload to Facebook button and will never post anything you don’t explicitly tell it to.
When posting to Facebook, you’ll be given an opportunity to include a Video Title and some text to go with the post. The Video Title is never really displayed on Facebook unless a visitor watches the video in full screen. The text component refers to the message that will accompany the video file, so be sure to include your commentary or call-to-action of some kind!
We’re building Lumen5 as a video creation platform with the goal of offering the world’s simplest video making experience, and our primary approach to this is to have the workflow assisted by AI. The idea is that the user would simply enter a link to a blog post or news article and Lumen5 will automatically turn the content into an engaging video. Lumen5 is designed for brands and businesses to produce video content at scale through automation.
The first bit of AI you’ll encounter is when you enter a link to a blog post or news article where our natural language processing algorithms will strive to help you summarize the article. As you know, a full article contains too much text for a short and engaging video, hence summarization is necessary. Currently, the system will detect and divide up various sections within the article and offer them as suggestions to you. There’s also a Smart Summarization feature which will help you summarize long run-on sentences. In the future, we hope to improve this algorithm further so that little to no human decisions are required, and all you get is a perfectly summarized storyboard for your video.
The next bit of AI you’ll encounter is our computer vision algorithms that automatically matches each of your scenes with corresponding background media such as photos, graphics, and videos. For example, if you entered an article about basketball, then the system will automatically seek out basketball-related media files and add them to your video. This area of intelligence will improve over time, allowing us to provide more and more accurate media matches and suggestions.
Many video details are also controlled by AI as well, such as the duration of each scene being dynamically determined based on the number of words in that scene. Computer vision is used to detect the focus of an image, such as human faces, and the system automatically positions the text to avoid covering the main subject of the background media. The system will also detect important phrases and keywords, then automatically applying a coloured highlight to bring attention to those words.
Last but not least, perhaps one of the most valuable component of Lumen5 is the built-in media library consisting of millions copyright-free photos and videos along with thousands of audio tracks. The toughest part to creating video content is securing all the assets you need to do it. Lumen5 does it all for you and you can search through the database with keywords then drag-and-drop them into your video along with an automatically generated credit scene at the end to ensure all attributions are properly in place.
The result is the ability to create engaging social videos within minutes rather than more traditional workflows which could take hours, or days!
The Lumen5 music library offers copyright-free audio tracks from the Creative Commons, allowing you to easily browse and add music to your videos. Previously, the music library offered approximately one hundred tracks.
In an effort to expand the size of our music library and improve the quality of your options, we have added a large collection of new copyright-free audio tracks. As a result, our music library has now been expanded from 100 to over 1000 tracks.
To accompany this expansion, we have redesigned the music library to offer a better user experience for browsing the new expanded library. This redesign introduces a new layout along with visible tags for each track. You can press the play icon to preview any of the tracks and use the new mood filter to browse through different music categories.
Lumen5 offers a number of branding options for customizing your videos. Previously, you can upload a single image file which is displayed as a small watermark in the top-right corner and also in the final scene of the video. Responding to user feedback, we have now separated these two options.
This new update will allow you to upload separate files for the watermark and the logo. There is also a new option to toggle each of them on or off individually, providing greater flexibility that previously available.
New scene types have been introduced in order to provide more styling options to your videos. Previously, the Text scene type was the only available option. With this new update, you can now also choose Title, Quote, or No Text.
Choosing Title will allow you to use a much larger font size for your text. The No Text option will remove all text from the scene, allowing viewers to focus on the image or video background. Last but not least, the new Quote option will allow you to stylishly include quotes into your video and also include an attribution for the quote as well.
Lumen5 is a video creation platform driven by A.I. designed to help marketers easily create high quality videos at scale through automation.
Videos get more likes, shares, and click-throughs. Posting videos on Facebook can 10x your reach.
Create videos in minutes
Our A.I. automatically creates videos for you. Just enter an article link and watch the video build itself.
Automate your videos
Add content sources and RSS feeds to have custom-made videos sent directly to your inbox on a daily basis.
Our A.I. platform helps you create videos in minutes. Here’s how Lumen5 works:
01 Enter an article link to turn it into a video
Using Natural Lanaguage Processing algorithms, our system will automatically create a storyboard for you.
02 Our A.I. will find the perfect audiovisuals
Using Computer Vision technology, our system will find relevant visuals and audio relevant to your content.
03 Add your branding and call-to-action
Upload your logo, add your watermark, customize your style and include a call-to-action to drive conversion.
04 Download and share the video
Videos on Facebook get 10x more reach. Platforms like Google and Twitter also rank videos above other types of posts.
Videos generated from company blogs using Lumen5