Publication by BgRem about creativity and design

BgRem API: Editing videos anywhere


BgRem API: Editing videos anywhere

We are introducing our new product extension for developers – BgRem API. It brings out a deep learning solution – Video Background Remover, developed by our ML team right into your application for rapid and easy integration. Bring a perfect ML experience to your customers with just two lines of code. 

The background removal model operates based on principles. This model suits our task of selecting all people in the frame and the objects they hold. It also helps when there are several people in the image. For purposes of training network, we have created our own segmentation dataset: the human image dataset developed by Deelvin. It takes into account short and long hair and all possible objects that people can wear and hold. You can read more about the segmentation of human images here

Semantic segmentation
Picture 1. Semantic Segmentation

The service consists of parts: the service is responsible for sending static files, processing user requests, uploading videos, forming a task queue, and workers taking the file and processing it. The service creates a queue of tasks that workers take in order to finish previous jobs.

Getting to know the service

BGRem API service allows you to upload and use backgrounds, both default and your own, and work with videos and images in a variety of formats. 

Picture 2. Background removal service

API is focused on automation of video and image background removing. With a few simple requests, you can get the same experience editing videos as on our website, look up documentation to see available commands, and get started.

To start using our service, first, you need to get an API key from our website; then, you can try out sample codes to test your integration. Our API works on any platform and in any system. Do not forget that each API call requires an Authorization header that contains your unique API key. If you are having troubles understanding responses or requests – refer to our documentation page.

Get Key GIF
Get key

Basic use steps API are:

  • Uploading video or image

Call POST /api/v1/videos with your file.

  • Fetching the video

Call GET /api/v1/videos/{video_id} to fetch the status of your video.

  • Downloading the result

When the status is donedownload the result from the result_url.

Our API supports the following file types for uploading:

  • Video files are supported in mp4, QuickTime, and WebM format.
  • Images are supported in jpeg and png formats.
API calls GIF
API calls

Look at our demo repository on GitHub to work with a sample project or to learn what calls you can perform with our API.

Result GIF


BgRem API is in development, and more features and options are coming. But you can try working with our API now already, and you don’t even have to pay – we have a trial plan which includes processing 30 videos having a 30-minute total duration. After it ends, you can continue using our service by subscribing to the Developer plan. It contains 340 minutes of video processing each month, and additional minutes are paid for after usage.If you have any questions or suggestions - we are always ready to help you or make improvements based on your feedback.