Releases
3.3.0#
Changelog#
-
AI Services: OpenVidu now supports a catalog of AI services that can be easily integrated into your application to enhance the user experience and add advanced features. These services are delivered through OpenVidu agents: a set of pre-configured and ready-to-use AI modules that seamlessly integrate into your Rooms.
We are starting with the Speech Processing agent: it focuses on transcribing audio speech to text and processing the results in various ways. Currently offering the Live Captions service, which generates live captions for your users' speech with great accuracy to display them in your frontend application. The Live Captions service supports many leading AI providers, such as OpenAI, Google, Azure, Amazon and more (see Supported AI providers).
Of course, you can also implement your own custom agents using the powerful LiveKit Agents framework and deploy it along your OpenVidu deployment. Any LiveKit agent is compatible with OpenVidu. Learn how to do so here.
-
Use a single domain for your deployment (EXPERIMENTAL): OpenVidu deployments now support TURN with TLS without an additional Domain Name using the flag
--experimental-turn-tls-with-main-domain. This is great for production deployments, as it allows you to use a single domain and still support users behind restrictive firewalls.You can deploy any OpenVidu deployment with this feature enabled:
- On Premises: perform a non-interactive installation passing the flag. How to perform a non-interactive installation for each OpenVidu deployment: OpenVidu Single Node COMMUNITY, OpenVidu Single Node PRO, OpenVidu Elastic, OpenVidu High Availability with DNS, OpenVidu High Availability with NLB.
- AWS: when deploying the CloudFormation template, add the flag
--experimental-turn-tls-with-main-domainto the parameter named(Optional) Additional Installer Flags", and leave empty parameters under(Optional) TURN server configuration with TLS. - Azure: when deploying the ARM template, add the flag
--experimental-turn-tls-with-main-domainto the parameter named(Optional) Additional Install Flags, and leave empty parameters under(Optional) TURN server configuration with TLS.
-
Azure deployment bug fixes:
- Media Nodes are now automatically deleted if the installation process fails, preventing unwanted resources being left in your Azure account.
- A misconfiguration was preventing the TURN server from working correctly in Azure. This is now fixed.
- Fixed a race condition during the deployment process in Azure that sometimes caused problems when creating multiple subnets concurrently (9728d96).
Version table#
3.2.0#
Changelog#
- OpenVidu Single Node PRO: OpenVidu Single Node PRO is a new type of OpenVidu deployment targeting users that want to deploy OpenVidu as a single-node setup but that still want the 2x performance boost and the advanced observability provided by multi-node OpenVidu PRO deployments.
- Azure deployments (Beta): OpenVidu now supports native deployments in Microsoft Azure. You can now deploy OpenVidu Single Node COMMUNITY, OpenVidu Single Node PRO, OpenVidu Elastic and OpenVidu High Availability in Azure using ARM templates. During version 3.2.0, Azure deployments will be considered in Beta.
-
New Azure recording tutorials: OpenVidu deployments in Azure use Azure Blob Storage to store recordings (instead of S3). For this reason, we have extended our recording tutorials with Azure Blob Storage compatible examples. You can find them in the following links:
-
External proxy configuration: By default, OpenVidu is deployed with an internal Caddy server to configure and manage SSL certificates. However, there are certain scenarios where using an external proxy might be preferable:
- You wish to manage SSL certificates manually.
- A specific proxy server is required for enhanced security.
- You need to integrate a proxy server already in your infrastructure.
For any of these cases, now all OpenVidu deployments allow configuring external proxies. You can find the instructions to do so in this how-to guide.
-
LiveKit stack updated to v1.8.4: OpenVidu 3.2.0 is now based on LiveKit v1.8.4, which includes several improvements and bug fixes. You can find the release notes here.
- OpenVidu installer improvements: Some users have reported issues when installing OpenVidu, which were finally caused by old versions of Docker and/or Docker Compose. The OpenVidu installer now checks both versions and displays a descriptive error message if they are incompatible.
-
OpenVidu Angular Components: see Angular Components documentation.
- Virtual Backgrounds improvements: More efficient use of resources by reusing the existing context. Avoid video flickering when changing the background. Improved resource reallocation management for smoother rendering. Contribution to LiveKit’s track-processors-js package (PR 86) resolving an issue affecting its dependencies.
- Fixed panel reopening issue with
ovAdditionalPanelsdirective. Custom panels created withovAdditionalPanelswould not reopen correctly after switching between default panels (activities, participants or chat). Now, returning to a custom panel restores it as expected without closing all panels. - Minor style fixes.
-
Deployment bug fixes:
- OpenVidu On Premises deployments that made use of v2compatibility module had a wrong configuration affecting the S3 MinIO bucket. This could cause issues when recording sessions from your OpenVidu v2 application. This is now fixed.
- Wrong Caddy configuration in OpenVidu High Availability deployments made some services not reachable in specific scenarios of fault tolerance. This is now fixed.
Breaking changes#
- For OpenVidu On Premises deployments, the default S3 bucket in MinIO has been renamed from
app-datatoopenvidu-appdata(in Single Node and Elastic deployments) and fromcluster-datatoopenvidu-clusterdata(in High Availability deployments). - Port rules in OpenVidu High Availability with Network Load Balancer have changed. Check the port rules from the installation instructions.
Version table#
3.1.0#
Changelog#
- IP cameras support: OpenVidu now allows you to connect RTSP IP cameras to your Rooms. This feature has been included in our custom fork of the Ingress module, which is used to ingest media into a Room. Check out how to do it here. IP cameras support has also been included into the v2 compatibility module. This means that if your OpenVidu 2 application is using the IP cameras feature, you can now upgrade your deployment to OpenVidu 3 and keep using this feature.
- OpenVidu Updater: you can now update the version of your OpenVidu deployment very easily using our new OpenVidu Updater module. OpenVidu Updater will take care of the whole process, from stopping the services to updating the configuration files. It will also manage backups to allow rollbacks in case of any issue. You can update your OpenVidu deployment from 3.0.0 to 3.1.0:
- Update your OpenVidu On Premises deployment: Update OpenVidu Single Node, Update OpenVidu Elastic, Update OpenVidu High Availability.
- Update your OpenVidu AWS deployment: for AWS deployment we recommend updating from 3.0.0 to 3.1.0 by redeploying the CloudFormation. From 3.1.0 onwards OpenVidu Updater will also be able to seamlessly update your AWS deployment.
- mediasoup stability: we believe we have reached the right point of maturity to take mediasoup as the internal RTC engine from experimental to production ready. There are still some limitations to take into account, but the general stability of the system is now considered production ready.
- v2 Compatibility bug fixes: there have been several improvements to the compatibility between OpenVidu v2 applications and OpenVidu v3 deployments:
- REST API: Field
clientDataof the Connection object wasn't being properly set. Now it is. - Webhook: webhook event
webrtcConnectionCreatedwasn't being sent when an audio-only Publisher published to the Session. Now it is. - openvidu-browser-v2compatibility:
- Event
videoElementCreatedwasn't being triggered for Subscriber participants. Now it is. - Event
streamCreatedwasn't being triggered by the Session object for Streams coming from audio-only Publishers. Now it is. - Event
streamPropertyChangedwasn't being triggered when an audio-only Publisher muted/unmuted its audio. Now it is.
- Event
- REST API: Field
Version table#
3.0.0#
Changelog#
- General Availability of OpenVidu 3, which is considered now stable and production-ready. Beta versions of OpenVidu 3 are preparing to be discontinued (including 3.0.0-beta1, 3.0.0-beta2 and 3.0.0-beta3).
Known limitations#
- When using mediasoup:
- No
ConnectionQualityChangedevent (LiveKit reference). - No
TrackStreamStateChangedevent (LiveKit reference). - Limited ingress support: non-simulcast video tracks are not supported. Firefox may experience issues when subscribing to ingress video.
- No
Version table#
3.0.0-beta3#
Changelog#
- Centralized configuration: OpenVidu now automatically manages and synchronizes the configuration of all its components. This means that updating any configuration parameter in multi-node deployments (OpenVidu Elastic and OpenVidu High Availability) is as simple as updating the required file in a single node. OpenVidu handles the distribution and restart of the affected services across all nodes. See how easily you can change the configuration here.
- mediasoup support:
- Dynacast is now supported (LiveKit reference).
- Adaptive Streaming is now supported (LiveKit reference).
- Speaker Detection events (LiveKit reference).
- Server API method
MutePublishTrack(LiveKit reference). - Client API method
RemoteTrackPublication.setEnabled(LiveKit JS reference).
- OpenVidu Call:
- When using it against an OpenVidu Local Deployment, recordings couldn't be accessed from the application's frontend. This is now fixed and OpenVidu Call is able to access recordings.
- There was an error when applying Virtual Backgrounds ("No camera tracks found. Cannot apply background"). This is now fixed.
- Docker image openvidu/openvidu-call is now 50% smaller.
- OpenVidu v2 compatibility:
- There was a race condition when multiple participants connected to the Session at the same time that could cause remote
streamCreatedevents to not be triggered. This is now fixed. - Configuration parameter
V2COMPAT_OPENVIDU_PRO_AWS_S3_BUCKETdid not allow configuring sub-buckets ("openvidu" worked fine, but "openvidu/subbucket" did not). Now it is possible to do so. - The operation to list recordings (available for REST API, openvidu-java-client, openvidu-node-client) was limited to 1000 recordings. This is now fixed and all recordings are always returned.
- There was a race condition when multiple participants connected to the Session at the same time that could cause remote
- AWS deployments: all secrets are now synchronized with AWS Secrets Manager. You can update any secret from the AWS console and restart your cluster for them to have immediate effect in all your nodes. This is also true in reverse: you can update any secret inside your node, and after restarting the cluster, the values in AWS Secrets Manager will be properly synchronized.
- New application tutorials available: iOS, Android, Recording.
Known limitations#
- When using mediasoup:
- No
ConnectionQualityChangedevent (LiveKit reference). - No
TrackStreamStateChangedevent (LiveKit reference).
- No
Version table#
3.0.0-beta2#
Changelog#
- Improved mediasoup support:
- Data messages work (LiveKit reference).
- Ingress supported (LiveKit reference).
- Improved OpenVidu Local Deployment:
- Fixed Room Composite Egress (LiveKit reference) support when using mediasoup.
- WebHooks (LiveKit reference) supported against a local OpenVidu Call.
- Production deployments have a better private IP discovery process when there are multiple valid private IPs in the same host. This will make more deployments work out-of-the-box without the need of manual intervention.
- OpenVidu PRO Evaluation Mode improved. Before, a maximum a 2 Rooms of 8 Participants each could be created. Now the upper limit of Participants still apply, but the number of Rooms is unlimited. For example you can have 4 Rooms of 2 Participants each, or 1 Room of 8 Participants.
- Minor bug fixes related to OpenVidu Call.
Known limitations#
- When using mediasoup:
- No support for Speaker Detection events (LiveKit reference).
- No
ConnectionQualityChangedevent (LiveKit reference). - No support for Dynacast (LiveKit reference).
- No support for Adaptive Streaming (LiveKit reference).
- When using OpenVidu Call against an OpenVidu Local Deployment, recordings cannot be accessed.