Do I need to write Python to use this workflow?

Only for step 4. Building the pipeline, configuring audit logic, running the HMI and reviewing results are all no-code.

Where do plugin scripts go?

Into ReliAudit\PostProcessed\plugins. Launch with main.exe and the script runs as part of the inspection flow.

What kind of logic belongs in a plugin?

Measurements, transformations and business rules. Thresholds and report formats belong in ReliAudit, where they can be changed without touching code.

How do images reach the runtime?

Through a folder monitor or a camera trigger, configured when you start runtime inspection in the web HMI.

How long does a first pass take?

Around 30 minutes once prerequisites are in place. Licence, GPU access and network routing usually take longest.

What changes when I move to a new product or line?

The pipeline or the plugin, not the platform.

How to Build and Execute an Inspection Pipeline with Custom Post-Processing

A 5-step workflow to take an AI inspection from pipeline design to production output — including the Python plugin layer where your domain logic lives.

5 steps
~30 minutes for a first pass

What You Will Build

By the end of this guide, you'll have a running inspection workflow that detects incoming images, applies your AI pipeline, and runs your own Python logic to produce production-ready output.

A Configured Pipeline

AI models and image-processing steps assembled visually in ReliVision.

Application-Specific Audit Logic

Rules, thresholds and report formats defined in ReliAudit.

Live Image Ingestion

A folder monitor or camera trigger streams data to the runtime live.

Custom Python Plugin

Your own post-processing — measurements, transforms, business rules.

ReliVision works almost like an SDK — keep the core inspection workflow, then plug in your own logic where it matters most.

Roadmap

The 5-Step Workflow

One connected flow from pipeline creation to production output. Each step is covered in full below.

Build Pipeline

ReliVision

Assemble image processing and AI models visually.

Configure Audit Logic

ReliAudit

Set rules, thresholds and report formats.

Launch & Ingest

Web HMI

Start runtime; feed images via folder or camera.

Add Post-Processing

Python plugin

Drop a script into the plugins folder.

Review & Verify

Audit review

Inspect outputs, annotations and structured data.

Time: about 30 minutes for a first pass once the prerequisites are in place.

Step 1 of 5 · ReliVision

Build the Pipeline in ReliVision

Build the inspection logic directly inside ReliVision.

Define how images will be processed and analyzed
Assemble inspection steps visually
Prepare the workflow for deployment

The pipeline defines how each image will be processed, analyzed and routed.

The Pipeline Editor with the blocks assembled, and the Import to ReliAudit dialog. — The pipeline is assembled visually, then handed over to ReliAudit for the application logic.

Step 2 of 5 · ReliAudit

Transfer to ReliAudit and Configure Audit Logic

Once the pipeline is ready, transfer it to ReliAudit and configure the application-specific logic.

Define what to inspect, including scope, assets and severity
Set rules and thresholds to classify findings and prioritize actions
Specify report formats, metrics and downstream deliverables

The ReliAudit audit list and the audit configuration panel with rules and thresholds. — Rules, thresholds and report formats are configured per audit rather than baked into the pipeline.

Step 3 of 5 · Web HMI

Launch the Web HMI and Start Image Ingestion

Select the audit, launch runtime and let images flow in for real-time inspection.

Select the relevant audit
Load and manage inspection processes
Start runtime inspection

Images can now flow in for real-time inspection.

The web HMI audit selector and the runtime view with a live inspected image. — The runtime HMI runs in a browser — no separate operator install on the line.

Step 4 of 5 · Python plugin

Add Custom Python Post-Processing

Use Python scripts to add calculations, transformations and application-specific outputs.

Navigate to ReliAudit\PostProcessed\plugins
Drop your Python script into the plugins folder
Launch with main.exe

The ReliAudit plugins folder beside the plugin script open in an editor. — A script dropped into the plugins folder is picked up automatically — there is no build or redeploy step.

Step 5 of 5 · Audit review

Review Outputs and Verify Success

Images are detected, inspected, post-processed and surfaced as operational results.

Detect incoming images
Trigger the inspection pipeline
Apply post-processing automatically

The audit review screen with inspected images, annotations and the post-processed output. — Annotations and post-processed values appear together, which is what makes the plugin output verifiable.

Why This Matters

Unified Inspection

Pipeline, audit, runtime and post-processing in one workflow — no glue code.

Adaptable at Scale

New product, new line, new rule — change the pipeline or plugin, not the platform.

Faster Deployment

Move from prototype to production without rebuilding for each application.

From pipeline to production output

ReliVision takes you from pipeline creation to fully customised inspection outputs in a single integrated flow — whether you are deploying standard inspections or building advanced application-specific logic.

Download ReliVision