VISUAL AI.
WITH SYNAHUB.
Ask your camera images a question. SynaHUB inspects visible conditions on a schedule or on demand, confirms relevant results and passes them into your workflows.
- Custom image questions
- Recurring inspections
- Appropriate follow-up actions


In brief
What is SynaHUB?
SynaHUB is Synaedge’s platform for automated visual inspections. It combines camera images with a specific question, schedule and suitable AI model. This creates repeatable inspection jobs: the platform evaluates images, checks results against your rules and passes relevant cases to your team or systems.
- Deployment
- Recurring inspections in production, logistics, buildings and other operational areas.
- Execution
- On a schedule, at intervals, on demand or triggered by a connected event.
- Operations
- With central job control and suitable compute resources on site, on your own servers or in a hybrid setup.
Questions from your operation
What would you like to
check in your images?
A good inspection starts with a specific question. These examples show how visible conditions can connect with your workflow.

Inspect clear routes systematically.
“Is the marked escape route free of obstacles?”
An inspection job evaluates the visible route at defined times. If a detected obstruction remains during a repeat inspection, your team can receive an alert.
Inspect → reconfirm → notify responsible team

Check the intended door state.
“Is the gate closed after operating hours?”
Combine the visible position of a door or gate with a time window. Confirmed deviations are documented with an image and timestamp for follow-up.
Time window → check door state → report deviation

Compare visible parts with the expected state.
“Are all required components in place?”
A reference image and a clear question describe the expected state. SynaHUB can inspect visible deviations and provide them for further review.
Define reference → compare image → document result

Detect visible replenishment needs.
“Is material still available at the designated staging area?”
Repeated image inspections can detect an empty or changed material location. A confirmed result can be passed into your existing replenishment process.
Set interval → inspect material location → trigger replenishment

Make the handover visually traceable.
“Is the delivery in the designated handover area?”
An on-demand inspection combines the current image with your operational question. The result, image and timestamp make the handover step traceable later.
Trigger job → inspect handover area → pass on case

Assess equipment at a defined inspection point.
“Are a helmet and high-visibility vest visible in the inspection image?”
Describe the visible equipment expected at a particular inspection point. We use your images to determine whether the angle of view and image quality are suitable.
Select inspection point → assess equipment → provide result
We use your images and desired outcome to determine which inspection is suitable for your task.
Combine as needed
Your inspections.
Your rules.
Combine a clear image question with the right time and follow-up process. You can combine these elements according to the task.
Inspection types
- Open image questions
- Reference comparison
- Reference-state inspection
- Visible completeness
Execution
- Schedules and intervals
- On demand or by event
- Camera groups
- Priorities and queues
Confirmation
- Repeated inspections
- Multiple confirmations
- Minimum duration
- Result rules
Handover
- Structured results
- HTTP interfaces
- Email actions
- Traceable history
From image question to action
One question.
A repeatable workflow.
You define what should be inspected and when a response is required. SynaHUB combines the image, model and your rules into a traceable inspection job.
01Define image and question
You choose the image source and describe which visible condition should be inspected.
02Run inspection
The job starts at the scheduled time. A suitable worker evaluates the image with the selected model.
03Confirm result
Your rules determine whether a repeat inspection or multiple confirmations are required.
04Trigger the next step
Confirmed results are documented and passed to your team or a connected system.
What belongs in an inspection profile?
The image source, question and timing form the starting point. Depending on the task, you can add an image area, reference image, priorities and confirmation rules. The expected result and next step are also defined.
How does a vision language model help?
A vision language model, or VLM, processes images together with natural-language questions. You can ask about an open door or missing components, for example. We use real examples to assess whether the answer is sufficient for your process.
How are results confirmed?
You can require several positive confirmations or a minimum duration before an action follows. Repeated inspections help distinguish temporary conditions. This does not replace an assessment of image quality and model suitability.
Platform insights
One interface.
Your visual inspections.
Explore eight original SynaHUB views, from inspection profiles and results to cameras, workers and models. Select a view and switch between light and dark mode.
Your inspection jobs at a glance.
View processed jobs, active workers and current alerts in one overview. Comparisons over time show how use of your platform is developing.


Define a question. Repeat the inspection.
Job profiles group your visual inspection jobs. Assign cameras and sites, set priorities and control which inspections are active.


Understand the result together with the image.
Read the inspection result for each zone and open the associated image crops. The history shows which condition was met and how the job was completed.


You decide when a result becomes an alert.
Connect inspection profiles with conditions and suitable follow-up actions. This defines which results reach your team and where they are passed on.


Assign your image sources centrally.
Manage cameras, customers and sites in one structure. This makes it clear which area an image comes from and which inspection belongs to it.


The right resources for your inspections.
Workers handle image evaluation. The central view shows their connection, capabilities and utilisation, allowing you to track how jobs are distributed.


What is queued, running or complete?
Track jobs directly on the worker. Priority, status and runtime help identify bottlenecks and monitor ongoing operation.


Choose the right model for the task.
Manage local models or connect an external VLM server. We use your images and expected results to determine which model is suitable.


Your infrastructure. Your processes.
Control centrally.
Process appropriately.
SynaHUB separates inspection-job control from execution. The Core manages tasks; workers handle image evaluation. Data paths and compute resources are adapted to your environment.
Your image sources
Cameras, camera groups and images from connected systems.
SynaHUB Core & workers
The Core schedules jobs. Suitable workers inspect the images with the appropriate model.
- Schedule
- Inspect
- Confirm
Your team & your systems
Documented results, alerts and appropriate follow-up processes.
What do the Core and workers do?
SynaHUB Core manages image sources, inspection profiles, schedules and rules. Workers execute inspection jobs with suitable models. This separation lets you allocate available compute resources according to sites and tasks.
Which models and operating locations are possible?
SynaHUB can connect local models and external VLM servers. We select the model and operating location according to image quality, task, speed and your data paths. Local processing and a coordinated hybrid architecture are possible.
How do results reach existing systems?
Confirmed results can be passed on through HTTP interfaces or email actions. Together, we define which data a target system needs and how the handover fits your process.
How can access and data be restricted?
Users, groups and API access control who works with the platform. Image scope, data paths and retention are defined for the use case. Which data leaves the site depends on the selected architecture.
We align image sources, models and interfaces with your existing operation.
Discuss integrationFrequently Asked Questions
Good to know before you start.
What is SynaHUB?
SynaHUB is Synaedge’s platform for automated visual inspections. It combines camera images with a specific question, schedule and suitable AI model. This creates repeatable inspection jobs: the platform evaluates images, checks results against your rules and passes relevant cases to your team or systems.
How does SynaHUB differ from SynaEYE?
SynaHUB carries out specific visual inspections as scheduled or triggered jobs. SynaEYE analyses live camera streams continuously and connects detections with operational events. The right platform depends on your task.
Can we use existing cameras?
Yes. Suitable cameras and existing image sources can be connected. We assess the technical connection, perspective and image quality. Images from a connected system can also serve as the basis for an inspection job.
Do we need to develop a custom AI model?
Not necessarily. A project often starts with an existing vision language model and a clearly worded question. A custom approach can be useful if existing models cannot solve the task adequately.
Can SynaHUB inspect individual images?
Yes. An inspection job can evaluate a single image from a camera or connected system. You define which question should be answered and how the result is used.
How reliable is a visual inspection?
This depends on the task, angle of view, lighting, image quality and model. We therefore evaluate the results using typical images and challenging examples from your operation. Rules for repeated inspections complement this assessment.
Do images need to be sent to a public cloud?
No. SynaHUB can run with local compute resources and your own servers. In a hybrid architecture, we agree which components run where and which data is transferred.
How do we get started with SynaHUB?
Describe a specific question and provide typical example images. Together, we define the expected result and next step. This creates an initial inspection profile that we test with your images.
Would you like to compare continuous video analytics and targeted inspection jobs?
Explore SynaEYE Compare platformsSynaHUB in your operation
Your images.
Your question. A first inspection job.
Show us what you want to inspect. Together, we select typical example images, define the expected result and test the first workflow.
