OpenTalk. Video conferencing designed from the ground up.

Recordings of conferences will soon be possible -- and can then be downloaded at the end in generally playable formats such as MP4 video files.

Also on the roadmap: Optional integration of chat and whiteboard into the recording.

Our chat feature allows users to send messages to single, multiple, or all participants of a session. This also applies to group discussions. Chat groups are set up automatically to help users avoid chatting to the wrong people by mistake.

Our roadmap:

Registered participants will be able to create their own contact lists for preferred chat partners.

Users will also be able to post questions to the moderator while running a group discussion. All participants can then vote up questions for the moderator or podium to see, to help them steer a session effectively and address those topics first that are most relevant to the audience.

OpenTalk adds some interesting participant views to the classical Grid view or Speaker view.

Different speakers can be highlighted for a limited time to allow for a more dynamic meeting experience. The video thumbnails of participants can also be displayed at a fixed position in the grid. That means everyone will know who "sits" where in the conference.

For podium discussions, a virtual stage will be created. This shows the video feeds of the podium speakers while hiding the other participants. The instructor in a virtual classroom is presented permanently, while the focus on individual pupils can be set to rotate, so that no one feels excluded.

The Dashboard presents important information in a structured fashion. This includes all scheduled meetings (your own and those you are invited to attend), clearly presented and arranged in a calendar view.

Recurring or important meetings can be highlighted and accessed through dashboard shortcuts. New meetings can be created instantly, using predefined templates.

During any video conference, it is possible to enter into a one-to-one audio chat with another participant by using the whisper key. The conversation won't be overheard by the people in the main session.

If a conference requires discussion to continue in smaller groups, OpenTalk will assist with the setup of break-out rooms and the assigning of participants. Rooms can be automatically created according to a range of different parameters. When group functions are used, the corresponding group rooms can be prepared in advance and participants assigned automatically.

Meanwhile, the conference moderators have everything under control: They can be called for assistance at any time and visit breakout rooms as required. A central event console gives access to all messages and advanced moderation features.

OpenTalk is purely browser-based and does not require any local software to be installed on the client. All common browsers such as Chrome / Chromium, Edge, Firefox or even Brave are supported. Functional limitations apply to Safari on Apple devices, as Apple does not support the WebRTC standard cleanly.

Centralized deployment reduces administration and operational overhead and enables ease of use for internal and external users. Depending on the authorization system, mechanisms such as single sign-on or 2FA are also supported.

Moderators can decide if participants will join a conference session automatically, or whether they need to be admitted manually from the lobby. The event console will tell moderators if there is anyone waiting to join.

Participants can confirm the performance of their current Internet connection and check and configure the functionality of their camera, microphone, or speaker in the lobby, before joining the actual video conference.

Participants can share individual program windows or their entire screen with the audience. The speaker video will remains visible when sharing. OpenTalk also supports parallel screen sharing by multiple participants.

Need to quickly discuss an issue privately with another participant? You can do so without losing touch of what is going on in the main session. Open a sub-conference room to have a private conversation but keep following events of the main event - in a separate window, with reduced audio volume and a video thumbnail image.

This is interesting for political debates, but also for teachers. For example, this feature could be used to help students during quiet, individual work without disturbing the other students in the class.

OpenTalk comes with a dedicated telephony feature - beside using the regular client over the Internet, participants can also dial in using a regular phone (audio only). OpenTalk can also be connected as a SIP client to an existing in-house telephone system or an external SIP provider. By doing so, conferences can be allocated a dedicated telephone number, and participants can dial in exclusively by phone or in combination with a computer that uses an Internet connection.

On our roadmap:

Future support for multiple phone numbers per conference, to offer dial-in worldwide. Registered participants to be automatically identified by their telephone number, and to have their name displayed accordingly.

For fun or privacy: A participant's background can be blurred or replaced with a virtual background image.

Background images can also feature a company's corporate design, provided on the server side,  and can be set as a default to be used by all company employees, if desired. Additional options for workgroup- or event-themed backgrounds to promote branding and increase immersion.

In addition to the common "Raise hand" gesture, OpenTalk supports many more features, such as choosing from different levels of urgency, which can be used to raise a point of order, or request topics to be added to the agenda, for example.

OpenTalk comes with speaker lists built in, so that moderators can manage individual speaker contributions as intended and give everyone the speaking time they are entitled to.

In virtual instruction, teachers can measure the engagement level in their classroom by inspecting and reordering the participant list to see the total speaking time for each participant, for example.

Our integrated whiteboard feature that allows for real-time collaboration with other call members in real-time, regardless of location. You can sketch out ideas, brainstorm, and make notes together on the same canvas. Multiple pen and color options are available, and it's easy to access within the tool. This promotes improved collaboration during meetings and more productive discussions.

Polls can be used to quickly gauge opinions, run a question and answer session in class, or as a way to add variety to activities.

OpenTalk offers polls where answer options include free text input, yes/no, or a selection of predefined answers. These can also be prepared in advance and presented as required to enable lively and hassle-free moderation.

In association and meeting law, motions to procedural rules or points of order must be given priority in a session.

OpenTalk offers features for lively debate and plenary culture, and support for motions. Such requests will be visible to the moderator and appropriately presented in the current list of speakers.

Together with our legally secure voting, this enables OpenTalk to facilitate virtual meetings for clubs or advisory boards, as well as political plenary debates.

Video conferencing can be fun: At school, university, or at work. Offering a variety of features is a breath of fresh air, keeps the suspense going, clears the mind and enables positive interaction between participants. Gamification is a central design aspect of OpenTalk. Examples are the "Wheel of names", theming, and the whisper button.

In a conference, short breaks may often last a bit longer than initially announced. Especially when all cameras are turned off, it's sometimes not clear to everyone when the session will actually continue, and five minutes may turn into ten.

In OpenTalk, moderators can set specific break times, turn off all cameras and microphones for the duration, if desired, and display a special break screen with a timer countdown. This way, everyone will know when the regular session is going to resume.

Every conference benefits from good time management. If conferences are time-limited, the remaining time can be emphasized with a color, and displayed in addition to the regular time (Like a "scrum clock"). Participants can be assigned a speaking time to avoid individual speakers to overrun their allocated timeslot - this also works with auto moderation and  talking stick activities.

OpenTalk supports multi-moderator meetings. Conferences, workshops, and plenary debates can be conducted more effectively if the effort of moderating these sessions is shared among several privileged moderators. For example, this would resemble the tasks carried out by a council of elders in parliament, or the main board members in a general assembly meeting.

An event console, which is visible to the moderator team, helps coordinate all pending tasks and messages. There is also a separate chat window for exclusive use by the moderation team.

Once a conference is over, the organizing team will sometimes perform a concluding evaluation, also called "off-boarding". OpenTalk supports such activities and if enabled, all presenters and VIPs can remain in a conference session after it has ended for regular participants.

Whether anonymously or not, as a quick way of gauging the mood in the room, or audit-proof for real commitment and impact: OpenTalk is ideal for data protection-compliant, secure decision-making. When it counts, for example at association, advisory board or political meetings.

Participants cannot vote more than once, and results cannot be manipulated by moderators. Even when voting anonymously, participants can use their "voting tokens" to double-check that their vote was counted correctly.

There is a debating technique where the person who holds the "talking stick" is the only one allowed to speak, until they pass it on to the next person. They can express their thoughts properly, and there is no risk of interruption while they have the talking stick. When finished, they nominate the next speaker by passing on the stick.

With this feature, even difficult or emotional debates with many participants can be conducted in a calm, structured, and constructive manner. It also promotes inclusiveness, as those who may find it more difficult than others to speak up are encouraged to offer their contribution.

Who is up next? With our "Wheel of names", you can leave it to chance to decide. A fun feature that can introduce excitement and entertainment in the classroom.

But the wheel of names can do more: If the objective is that no one is to be left out of the conversation, the wheel can ensure that those who have already spoken will not be called up again.

Effective team building: Any experienced team seminar leader knows the so-called "ball of yarn" activity, which is often used for welcoming and introducing participants who do not know each other yet.

After offering their own introduction, the current speaker will hold on to the string and throw the ball of yarn towards another person in the meeting, whose time it is to speak - over time, this will create a web of yarn between all participants. When the last person has had their say and introductions are finished, everyone ends up connected to everyone else.

For improved meeting control and time saving, OpenTalk includes master moderator features such as "mute all participants". This allows moderators to mute all participants at the same time with one click – or even or select individual participants. It helps to increases productivity by eliminating background noise.

Our protocol feature allows the moderator to set one or multiple participants for note taking and saving them in PDF format. This saves time and streamlines processes, increasing productivity.

"Hands down" is another moderator feature that allows the moderator  to prompt participants to lower their hands and move on to the next topic. This promotes engagement and gives moderators greater control over the meeting, making it easier to manage and stay on track.

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.

OpenTalk supports customization of logo and colors to fit your corporte identity. Presets and features can be configured system-wide, client-by-client, or at an individual level, per conference. Giving you the flexibility to brand OpenTalk to suit your needs.

On our roadmap:

Provide background images from the server end to fit your corporate design, to be used by everyone affiliated by default.

OpenTalk supports the operation of conferences in a cluster of federated conference servers, which enables the creation of security zones. A new conference can be bound to a specific conference server, which hosts the meeting. Participants need to access this specific server to attend the meeting. This way, some conferences can be accessed only from pre-defined locations, such as the computers of your in-house network. Thus, any security-sensitive content will be protected by an additional level of security.

On the other hand, open conferences with external participants can be held in a way that optimizes traffic, using video bridge servers that are freely accessible over the Internet.

With OpenTalk, each client will send audio and video streams in three different resolutions simultaneously. This allows other participants to receive this content in a way that is optimal for their particular setup and location. Each client performs a continuous bandwidth check and will swap streams adaptively to provide the best-possible user experience for everyone.

Deploy OpenTalk in a way that best fits your needs and circumstances. For total digital sovereignty, OpenTalk can be operated locally, on your own servers, and connected to your local administration interface. Alternatively, get access to OpenTalk through our software-as-a-service (SaaS) arrangement, where we take care of the operation of your conferences. You can rely on the scalable resources and bandwidth available to our video servers, and the fact that we will always keep the software up to date, and run the system within a hardened security environment.

A third option is a mixed mode operation, where we provide the control and authentication of OpenTalk as a service, and a video bridge in your own network is then used for running the actual conference. This keeps network traffic and bandwidth parameters at the local level and ensures that any security-sensitive conference data never leaves your premises.

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.

OpenTalk offers excellent image quality even on low-bandwidth Internet connections, using HTML5 and video codecs that are optimized for the Web. For corporate use, that means resource consumption and the additional load that is put on your company's infrastructure is kept low.

OpenTalk's special multiplexing technology ensures best-possible bi-directional audio and image quality for all participants, even in situations where bandwidth is sketchy. Whenever available bandwidth is scarce or traffic expensive (e.g., data roaming), OpenTalk can be set up to operate in an "audio-only" mode.

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.

Security experts praise the programming language Rust. For OpenTalk, Rust was an obvious choice not just for its security features but also because it lends itself well to the development of software that does a lot of parallel processing, as is required for video conferences with high numbers of participants. The Rust compiler detects programming issues that can lead to malicious memory access and buffer overflows, and provides protection against modern attack vectors out of the box.

For OpenTalk, a completely new architecture was developed from scratch that offers the latest in security, authentication, encryption, scalability, and flexibility: A state-of-the-art architecture from the 2020s.

Since clients subscribe and receive only those video streams that are currently played back, the feeds of those participants that are not displayed have only negligible impact on client scale-out in terms of computing resources and bandwidth. On the server side, they generate incoming bandwidth, but computing-wise, OpenTalk scales very efficiently since there is no image processing required on the server side.

Since any number of video bridges can be used, and each conference is launched in its own container on a horizontally scalable Kubernetes infrastructure, large scale-out will only be limited by the provision of further Kubernetes nodes by the operator. There is support for utilizing cloud providers for this, and these can be automatically connected.

A large number of (passive) viewers can be reached by connecting streaming platforms such as NC3 or Youtube. This is planned for an upcoming release.

Video streams can be exported to external streaming providers. Primarily, RTMP(S) is used, which is supported by Youtube and many other streaming services. As an alternative to exporting to an external service, OpenTalk also provides a dedicated light-weight streaming server that can be accessed through a web browser. All available authentication options can be implemented, if required. On request, features such as a "chat board", or "question board" can be implemented for external participants, while passive stream participants can submit questions to the moderator outside of the "official" stream of chat messages.

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.