Skip to content


Komposition is a highly opinionated piece of software, specialized at editing screencasts. As such, it prescribes a specific recording and editing workflow. To reach optimal results and have a nice experience doing so, please follow the guidelines outlined below.


It all begins with the recording. For Komposition to be able to automatically classify the parts of your screencast, and for you to be able to compose those parts in a fine-grained and effortless manner, your recording needs to be done in a certain way.

Write a detailed script before recording anything.
This is paramount for the Komposition workflow. Also, it's generally good advice that will save you loads of time, even if you're not using Komposition. Your script should be detailed enough for you to record video and audio separately based off it.
Record video and audio separately.
Video and audio recorded simultaneously is hard to separate and work with independently. The microphone easily picks up the sound of you typing or clicking the mouse, making the audio track inherently tied to the video track. If you're not a native English speaker, recording Finally, it's easier to get a clean audio track without computer fans humming if you record it separately.
Take long breaks, two seconds or more, between the parts in your screencast.
For the scene and sentence classifiers to pick up the different segments, you need to take long explicit breaks. A "part" of a screencast should be a short sequence of actions, and the corresponding sentence, or couple of sentences, describing that action. The more fine-grained you make your parts, the more control and comfort you'll have when editing the screencast in Komposition.

Composing Parts

Now that you have recored your video and audio, you can import using automatic classification. This will give you a library loaded with fine-grained parts of video and audio. Using the commands you compose your parts in sequences and parallels into a complete screencast.