Stronger together

OpenTalk comes with open and well-documented APIs. Clearly specified interfaces enable seamless integration into existing platforms and products. All access to features is subject to strict authentication and authorization.

REST-based APIs are available for provisioning of users and rooms, configuration, metrics, and reporting.

This enables a rich set of use cases beyond pure video conferencing and makes OpenTalk a versatile video communication solution for third-party applications.

OpenTalk utilizes modern container technology. Each conference will be launched within a container. For small-scale installations, that's usually "localhost", or alternatively, a Kubernetes cluster for anything bigger. OpenTalk takes care of the provisioning, operation and shutdown of each conference container.

Containerization ensures vertical & horizontal scaling as wells as clear data separation between conferences and participants. OpenTalk allows hassle-free operation, good management of CPUs and other resources, and avoids the "10K" scale-out issues that other solutions may suffer from. A conference container runs largely independently, which is great for administrators, as configuration changes and even system updates can be performed at all times without causing disruption.

The visual design of OpenTalk can be customized for specific projects. This includes an individually designed login page and the central provision of additional backgrounds. Further requirements can be checked by the partner at project level.

Roadmap: In future, it should be possible to adapt the color scheme to your own CI or add your own logo based on simple configuration settings.

OpenTalk is made for providers. Clear interfaces allow for clean integration, straightforward management of permissions, and flexible configuration options per customer or conference. This includes flexible billing options like basic, premium and enterprise packages as well as reporting.

Extensive metrics that can be read out enable precise recording of the scope of use and operating costs, very flexible configuration options of OpenTalk support the operation of the most diverse low-cost models and tariff groups with individually packaged functional scope.

Using well-defined interfaces and APIs, all kinds of qualitative and quantitative metrics can be retrieved from the OpenTalk platform for monitoring and performance analytics. 

OpenTalk allows you to stay on top of all mission-critical indicators: Server health, available resources and consumption by container, participant numbers per conference, number of connections, reconnects, other events, usage time, and many more.

Good reporting and meaningful statistics show the current status and predicts future situations, so that impending bottlenecks can be identified at an early stage and available resources allocated accordingly.

OpenTalk is based on an architecture that has been developed from scratch and that reflects the current state of the art in security, authentication, encryption, scalability and flexibility.

A central "controller" supervises the logon and authentication of all users and confirms their authorization to access sessions. After the initial authentication has been successful, another authorization check is performed on the basis of an OpenID Connect token, for each access: Whether it is retrieving video streams from an ongoing conference or transmitting required control information, chat messages, voting participation or using any other feature within the scope of participation. Nothing takes place without token-based authorization.

Only after logging in will the client connection be internally forwarded to the specific video bridge from which the respective conference content is accessible. The bridge will only ever receive authorized RTC connections, which ensures that only trustworthy and authorized audio/video data are being disseminated. The internal systems can thus be protected and hardened against unauthorized access and denial-of-service attacks by way of a multi-level system ("onion skins").

Control commands, such as microphone on/off and camera on/off, are first authorized by the user client and sent to the controller, which then forwards them to the responsible video bridge with the help of a so-called message broker. The message broker has been designed for high-performance scaling and can also simultaneously process the message volumes that occur in those conferences that have a very high number of participants.

For each new conference, a separate video bridge instance is started in a container, which processes all security-sensitive data locally within the container. This protects against unauthorized access from other installations/conferences and, by dismantling the conference container upon termination, also ensures that all data stored at runtime is securely and reliably deleted at the end of the conference.

in terms of streaming quality, modern standards and license-free open source video codecs such as VP8, VP9 or AV1 are efficient when it comes to data consumption and facilitate a good, low-latency conference experience, even with limited or fluctuating available network bandwidths. The connection quality is continuously monitored and the connection parameters are dynamically adjusted to the available bandwidth.

The frontend or user interface of OpenTalk is based on React, an efficient JavaScript library that works with hierarchical components. To guard against cross-site scripting (XSS) attacks, React automatically masks all HTML data entered by users. The Redux library ensures that runtime data is only stored locally unless needed by other users, allowing for a data-efficient operation.