Help your VLM understand the physical world.
Vivu helps robotics teams improve how their vision-language models recognize real-world objects and understand the relationships between them. We focus on the visual understanding your application needs: which objects are present, and how they relate to one another.
Recognize the objects that matter.
Improve how your VLM identifies the objects relevant to your robotics application, with a focus on the scenes and environments your team works with.
Understand how objects relate.
Go beyond identifying individual objects. Improve how your VLM interprets relationships within a scene, such as an object being on a surface, inside a container, or next to another object.
Put your recorded robot runs to work, too.
Search operational recordings in plain language and retrieve candidate moments with source videos and timecodes. Use the footage for incident review and for human-reviewed training, evaluation, and QA datasets.
The Free plan covers hosted search of recorded video.
Upload only video you are authorized to share and process. See our Terms and Privacy Policy.
Evidence instead of scrubbing
Describe what happened and get candidate moments back, each carrying the recording it came from and its timecode. The review starts from the event rather than from a file name.
The review outlives the reviewer
The index belongs to the team. When the person who knew which recordings mattered moves on, the ability to find things in the archive stays behind.
One incident becomes a pattern
The same kind of moment can be pulled across separate incidents, so a single reconstruction turns into material for QA and for training remote operators.
What does Vivu optimize for robotics teams?
We focus on how VLMs recognize objects in real scenes and understand relationships between those objects. The work is scoped to your model, visual tasks, and operating environment.
Can we also search our recorded robot runs?
Yes. Upload recordings into a Vivu project and search them in plain language once indexing finishes. Results are candidate moments with source recordings and timecodes, ready for your team to review.
How do camera recordings in robot log formats work?
For hosted video search, camera streams stored in formats such as ROS bag or MCAP need to be exported to video files before upload.
Further reading
All pagesBring us the scenes your VLM needs to understand.
Tell us about your model, the objects it needs to recognize, and the relationships it needs to interpret. Let's discuss where Vivu can help.