Cloud Comparison · Private Beta

Keep raw logs local.
Compare approved derived Bundles in VeluneOS.

Use derived Reference and Target Comparison Bundles generated in your environment to structure observed differences and identify where engineers should look first.

Apply for Private Beta View how it works
Raw log stays local Upload preview before transfer Derived Bundle only Human final judgment

Who It Is For

We are validating with these teams first.

This Beta is not a public SaaS supporting every robot log format. We are starting with a small number of Design Partners that have a clear Reference–Target comparison need.

Teams comparing before and after

Teams that need to identify which observations and investigation windows changed after software, configuration, or sensor updates

Teams manually reviewing repeated runs

Teams spending time replaying Reference and Target logs side by side to locate changes manually

Teams that cannot transfer raw logs

Teams that must keep raw logs internal while allowing only limited derived structures to be processed externally

How Cloud Comparison Works

Prepare with the Adapter. Compare in the cloud.

Download the VeluneOS Adapter and create comparison Bundles from your raw logs in your own environment. Upload the Reference and Target Bundles, and VeluneOS turns the differences into a comparison report.

STEP 01

Download the Adapter

Download and install the VeluneOS Adapter that matches your current log format.

STEP 02

Create comparison Bundles

Keep raw logs local and create the Reference and Target Bundles in your environment.

STEP 03

Upload the Bundles

Select the two Bundles and upload them to Cloud Comparison.

STEP 04

Review the comparison report

See what changed between the two Runs and where engineers should look first.

Data Boundary

See what may be transferred and what stays local.

COMPARISON BUNDLE

Derived structures that may be transferred after approval

  • Canonical Topic Role
  • Relative-time events and Evidence Windows
  • Gap, jitter, and count profiles
  • Structural features required for Reference–Target comparison
  • Context explicitly approved by the customer
STAYS LOCAL

Raw data not accepted by standard Cloud Comparison

  • Raw MCAP and rosbag2
  • Raw images, video, and audio
  • Raw point clouds and actual GPS
  • Actual topic and node names and absolute timestamps
  • Direct company, site, and robot identifiers

Exporter output is not automatically described as fully anonymous. During the initial Beta, Bundle structure and transfer boundaries are reviewed together with the customer.

Cloud Comparison

Cloud Comparison. Free or paid.

The same Cloud Comparison. Choose how your data is used.

FREE Free
₩0

How Cloud Comparison Works

01 Download the Adapter
02 Create Reference and Target Bundles
03 Upload derived Bundles
04 Review the Cloud Comparison report

Uploaded derived data is used only for report generation and VeluneOS improvement.

Start Free

Beta Operation

We validate the operating model first, then automate it.

Initial Private Beta

The Comparison Engine generates the report while a VeluneOS operator reviews input errors and execution status in a service-led operating model.

  • Pre-check Bundle schema and support eligibility
  • Handle Comparison Job failures
  • Review the quality of engine-generated reports
  • Collect engineer feedback
Automated operating model

The service evolves so customers upload Bundles directly and the server automates validation, comparison, reporting, feedback, and deletion.

  • Automated Reference–Target comparison
  • Report URL, JSON, and Markdown delivery
  • Organization access and retention controls
  • Operators handle exception jobs only

Design Partner Application

Apply for Cloud Comparison Beta

This information is used to assess Beta eligibility and supported log structures. Do not enter raw logs, access keys, personal data, or sensitive incident details in this application.

Back to main page

The button opens an application email in your default email client. This page does not store or transmit the entered information to a server.