Intelligent Video Solutions has released VALT 6.7, continuing VALT's localized AI development journey and laying the foundation for the next wave of features coming to the platform. Alongside those headlines, 6.7 introduces two entirely new features, Playlists and Freehand Draw Comments, along with a long list of improvements to how teams review, share, and manage video every day.
Here's what's new, and what it means for the people using VALT day to day.
AI isn't new to VALT. Our automatic keyword recognition, motion-activated recording and the addition of markers (annotations) to a recording based on key spoken words focused on solving real problems for real workflows.
Most programs already have strong, highly customized methodologies. They don't need platform technology defining how they teach, assess, or run their center. AI systems need to be trained on real data to be useful, our belief is customers should be empowered to own and decide how that data is used.
That belief shapes our priorities: Provide secure locally hosted AI environments trained from in-house pedagogies that remove friction from video review and debrief, supports educators rather than replacing their judgment, and AI developed at the edge wherever possible.
Local transcription in VALT 6.7 builds on this principle with automatic, on-premises transcription, extending VALT's locally hosted AI capabilities. Recordings can now be transcribed automatically as soon as they finish, or on demand from within Review, without any data or audio ever leaving the customer's network. For teams in HIPAA-regulated or air-gapped environments, that's the difference between being able to use AI transcription at all and not.
From a user's perspective, local transcription just works the way you'd expect: transcripts can be edited in place, exported, and searched, and administrators can set up entire user groups to have every new recording transcribed automatically. It's also flexible on the back end, running independently per GPU server, so organizations already using AWS transcription can run alongside a local transcription on others. Nothing is forced, and existing workflows aren't disrupted.
View GPU-based AI-compatible VALT server appliances >>
Before anything else shipped in this release, VALT went through a full third-party penetration test, a formal, coordinated security audit and remediation process. For organizations in healthcare, law enforcement, and government, that means VALT now comes with documented, third-party-validated security testing you can point to directly in RFPs, InfoSec questionnaires, and compliance reviews.
That security work extends into day-to-day access controls as well. Administrators now have far more flexible single sign-on configuration, with support for complex, multi-attribute group mapping rules. Motion detection can be turned on or off per camera rather than for an entire room, and sharing links now offer the same granular toggles for comments and guest markers that already existed elsewhere in the platform. Together, these changes give administrators tighter, more precise control over who can see and do what.
Playlists are a brand-new feature area in 6.7, and they solve a problem a lot of VALT customers have had for a while: how do you bundle multiple recordings together into a single, ordered viewing experience for training, review, or instruction?
With Playlists, users can select any recordings they have access to, and combine them into a curated set that plays continuously, one after another, with easy sidebar navigation between clips. Playlists can be renamed, reordered with drag-and-drop, and shared internally with a direct link, so recipients get access without needing to track down each individual recording. Deleting a playlist never deletes the underlying recordings, so there's no risk in experimenting with how you organize content.
This is especially useful for training coordinators building onboarding libraries, coaches assembling highlight reels for review, and educators sequencing case studies for students to work through in order.
The single biggest feature investment in 6.7 is Freehand Draw Comments, a full drawing toolkit that lets reviewers annotate directly on top of the video itself. Users can draw freehand, add shapes and straight lines, insert text, choose colors and line widths, and erase or undo as needed, all overlaid on the video during the moment they're commenting on.
These drawing markers behave just like any other comment type in VALT: they can be replied to, edited, deleted, and searched. For anyone who has ever tried to describe "look at the positioning in the top-left corner around the 2:15 mark" in a text comment, this feature makes that kind of feedback immediate and visual, which is especially valuable for coaching, clinical simulation debriefs, and skills evaluation.
A number of smaller updates in 6.7 make day-to-day review noticeably easier. Loop playback brings continuous playback to both individual recordings and full playlists with a single toggle, no admin setup required, which is ideal for training kiosks or stations that need to replay content on repeat. PTZ camera control preferences (joystick or compass style) are now set per user instead of as one global setting, which matters for shared workstations and simulation labs where different people prefer different control styles. Recordings are displayed with icons to differentiate types of media.
On the observation side, WebRTC streaming has moved into the default streaming protocol for new installations, delivering ultra-low latency for live viewing, and new 12- and 16-camera grid layouts let control rooms and simulation labs monitor more feeds without opening additional windows. Observers can share comments, add attachments, and allow guests to create markers and see the live and paused state of the video, creating more cohesive review across teams.
For evaluators, a new multipoint matrix field type supports rubric-style grading with scored selections across rows and columns, and multi-camera recordings can now be downloaded as a single lossless zip file containing every camera angle.
A handful of updates also give administrators more control with less manual work. Bulk actions let administrators share a whole batch of recordings at once or reassign the author on many recordings in a single step, saving significant time during onboarding or staff transitions. Schedules can now be exported individually or in a selected group, creating useful audit trails, and a new overlap warning flags conflicting schedules before they cause recording issues.
Finally, a few quality-of-life improvements round out the release: text fields in templates allow up to 4,096 characters (up from 1,024), scroll bars are wider and easier to grab, and the delay before a camera is flagged as offline is now configurable, reducing false alerts from a flickering feed.
VALT 6.7 continues to build the on-premises, GPU-powered foundation for the AI capabilities planned for upcoming releases, guided by the same priorities: reducing friction, supporting educators, and keeping data in the customer's hands.
IVS is focusing on actively researching and developing localized AI models that run entirely on your own VALT server, audio and video analytics that make use of camera hardware already in place, and customizable analytic engines that align to a center's specific use cases, all while keeping sensitive data in-house rather than sending it back to IVS or a third-party provider. Features like OCR that reads on-screen patient vitals and manikin displays directly from video, automated session summaries, and comment intelligence are early examples of where this is headed.
VALT is moving toward a future where the platform does more of the heavy lifting of reviewing and understanding video, while allowing users to focus on training, educating, and research.
Contact us if you would like to upgrade your on-premises server capabilities.