pixo is able to process a wide range of data, including both structured and unstructured formats. For example, it supports common file types such as PDF, XML, JSON, Excel, CSV, Word and text files.
In addition, data can be received in real time via REST APIs. This means that pixo is always up to date.

pixo can be seamlessly integrated into existing IT infrastructures by providing a REST API. This allows us to easily connect pixo to your preferred communication platform. In addition, a number of ready-made interfaces for common applications are already available.

pixo attaches great importance to the security of your data and is fully GDPR-compliant. The underlying AI model is selected in close consultation with you to ensure that it meets data protection requirements.
In addition, pixo offers the option of a complete on-premise solution that can be operated without internet access. Maximum data security and autonomy are thus guaranteed.

Your data is stored securely in your pixo instance. They are stored both locally as files and in vectorized form in a specialized database.

Yes, pixo can be operated both on-premise and in the cloud to meet the different needs of companies.

Advantages of the cloud solution

• More cost-effective, as less infrastructure investment is required
• Easier setup and less maintenance.

Advantages of the on-premise solution

• Higher data security (especially important for sensitive data).
• Operation without an internet connection is possible, which offers additional security
• Complete control over the IT infrastructure, but at a higher cost.

Mixed solutions

Mixed solutions where, for example, data and pixo are operated on-premise while the AI model runs in the cloud are also possible. This offers flexible adaptation to specific security and performance requirements.

The layered architecture of pixo is designed to enable a clear separation of the essential components of the platform. This includes storing the data as pure files, converting it into a vectorized representation for more efficient processing and analysis, handling the AI model separately, and connecting it to your preferred communication platform via REST API.

The advantages of this architecture are

• Flexibility to integrate new technologies: Due to the separate layers, new AI technologies can be quickly and easily integrated into the existing system, allowing pixo to always keep pace with rapid technological progress.
• Minimal maintenance and downtime: The independence of the layers from each other reduces the complexity of maintenance work. If updates or changes are made to one layer, the other layers remain unaffected, which increases system stability and minimizes downtime.
• Adaptability and scalability: The modular structure allows individual components to be scaled or adapted independently of each other, depending on the current requirements and loads on the system.
• Security and data protection: Each layer can be individually secured, which enables fine-grained control over data protection and access rights.

pixo ensures the accuracy and relevance of the information provided by using state-of-the-art techniques and algorithms that are continuously updated and expanded. The system enriches the original queries with the most relevant information from the data sets before sending them to the AI model. This provides very precise answers. However, it should be noted that 100% accuracy cannot be guaranteed when using any AI technology, as this is not yet technically feasible.

pixo can use most LLM AI models and is not tied to a specific model or manufacturer. We support almost all Large Language Models (LLM), including popular options such as GPT-3.5, GPT-4, LLama 2 and Mixtral 7x8b. Thanks to our layered architecture, AI models can be swapped as needed to always use the best model available. This enables us to react quickly to changes in regulations. For more detailed information on the specific AI models, please contact us directly. When introducing pixo, we will suggest the best model for your individual use case.

To provide the best possible support for pixo users, we offer a wide range of training and information materials. These resources are designed to give you and your employees a comprehensive insight into how pixo works and how to use it. In addition, we offer sensitivity training and guidelines to ensure that your employees are not intimidated by the use of AI. These measures serve to promote the acceptance and effective use of pixo in everyday working life and to develop a positive relationship with the technology. We will develop individual measures for your company together in a workshop.

The LINXYS team is at your disposal for technical support. We always respond to inquiries as quickly as possible; you can generally expect a response within 24 hours on working days. The exact conditions and details for technical support are discussed in a joint workshop, in which we address your specific needs and requirements. There you can also ask questions directly and obtain further information on support services.

You can use pixo in two ways: Firstly, via the free web interface, which offers a simple and intuitive user interface. Here you can make requests directly, view data and use the platform’s functions.
Secondly, pixo can be seamlessly integrated into your preferred communication platform, such as Microsoft Teams or Slack. This allows you and your team to use pixo’s functions directly in your familiar working environment.

Yes, with pixo, individual access concepts can be created that regulate access to data depending on employee role or department. For example, access for sales employees can be restricted to certain data, while developers have access to a different selection of data. These access permissions are developed in close cooperation with your company during a workshop to ensure that the data access rights meet your company’s specific requirements and security guidelines.

A pixo update can be carried out on two main levels: Data updates and software updates.

Data updates
Data updates in pixo can be performed either at set intervals or immediately, depending on your organization’s requirements. These updates usually have a very short downtime of around 1 to 5 minutes. A common approach is to perform the data updates at times of low activity, for example at night or at weekends, to allow the learning phase of the new data with minimal interruption.

Software updates
The way in which software updates are carried out depends heavily on the hosting setup, whether on-premise, in the cloud or a hybrid solution.