SINAPSI: How your video archives can become an e-learning portal thanks to Artificial Intelligence

Sinapsi is an IT company specialising in software for construction companies, especially those that provide specialised personnel to support emergencies, such as heavy vehicle repairs and maintenance. Founded in 2000 as an offshoot of Imola Gru – specialised in crane repairs and maintenance – with the specific purpose of improving company management through software. For Imola Gru, Sinapsi created Logica, a management software that then took on a life of its own and is now sold to many other companies.

The challenge

The client has a large archive of video guides, which are divided into many parts and are difficult to sort and search. To solve this problem, the client turned to Claranet, requesting an advanced tool capable of processing the videos (merging them, generating subtitles, etc.) and allowing efficient keyword searches.

The solution

Claranet has designed and developed a sophisticated platform that combines artificial intelligence, machine learning, and cloud technology for video cataloging.

The video processing is based on several cutting-edge components and technologies

Video Processing with FFMpeg: The system uses FFMpeg for video processing, allowing you to merge, compress and optimise content for smooth playback and low impact on system resources.
Asynchronous Processing Pipeline: Processing is done asynchronously to ensure high performance and reduce waiting times. The user can select the videos and the desired operations (compression, subtitle generation, noise removal, voice isolation, transcription dubbing, subtitle translation), starting a job that automatically manages the various steps of the process.
Integration with Amazon Web Services (AWS): The platform uses AWS services such as Amazon S3 and Amazon Transcribe for video storage, transcription and indexing. This allows us to offer a scalable, secure and reliable service.
Process Status Monitoring: The user has a dedicated page available to monitor the status of the processes, obtain information on any errors and access the system logs.
Using Artificial Intelligence for Cataloging: The system leverages the Anthropic Claude 3 model on AWS Bedrock to analyse video transcripts and generate relevant labels that facilitate intelligent categorisation of content.
Advanced Search with Elasticsearch: Video metadata, transcript, and model-generated labels are saved in an Elasticsearch database, enabling a fast and intuitive search experience, thanks to the features offered by this powerful search engine.

Thanks to this advanced and customised solution, the customer can now manage their archive of video guides in a simple and effective way, significantly improving the organisation and usability of the contents.

Next steps

In the next steps, we will focus on extending access to the platform so that it can be used by Sinapsi customers, while ensuring the security of data and information with the adoption of cutting-edge protection protocols and technologies. In addition, the platform will be enriched with new video processing capabilities to further expand the range of options available to users.

Find out more about Claranet's Data and AI solutions, and speak to one of our experts today.