On-Premises First
Customer-controlled deployment is our target architecture.
For automotive embedded software teams
DebugScout is an AI-assisted investigation platform for automotive software integration and support teams. It helps engineers narrow down where behavior diverged, with every finding tied to evidence and reviewed by an engineer.
See how DebugScout analyzed three sample AUTOSAR error scenarios.
First workflow shown: reference vs. faulty log comparison · simulated demo with synthetic data · no sign-up

The problem
Some support investigations begin with only a diagnostic trace and no customer debugger access. Authorized supplier engineering teams may also have source repositories, configuration files and software revisions. Connecting that evidence to a customer fault report still requires careful engineering investigation.
Hours per ticket
Manual log review dominates investigation time.
Different access boundaries
Some teams work from logs alone; others have authorized repository and configuration access.
Knowledge silos
Expertise in state-machine behavior sits with a few senior engineers.
The approach
The first workflow we demonstrate is log comparison. DebugScout is also being developed for authorized supplier engineering teams with access to source repositories, configuration files and customer diagnostic logs. The broader product vision is source-aware, evidence-cited fault investigation, from first symptom to an engineer-reviewed conclusion.
Select logs
Pick a known-good reference run and the faulty customer log.
Compare behavior
State transitions are aligned side by side. Missing and unexpected events are highlighted.
Review findings
Each finding cites the exact log lines behind it and lists possible explanations, not a confirmed root cause.
Engineer decides
An engineer approves or rejects the summary. Human review stays in control of every conclusion.
Walk through a synthetic body-controller wake-up fault where an expected state transition is missing from the faulty trace. Preloaded logs, precomputed analysis, clearly labelled as a simulation.
Security & Deployment
DebugScout is being designed for engineering teams working with sensitive source code, diagnostic logs, configurations, and customer information. Security and deployment requirements are central to our product development.
Customer-controlled deployment is our target architecture.
Authorized repository access and project isolation are design priorities.
Local inference is planned, with clear controls around any external AI services.
Defined retention and deletion policies are planned for pilot deployments.
Designed to support deployment inside customer-controlled engineering environments, with source code and logs processed according to agreed security requirements.
Potential future deployment option, subject to technical validation, security review and customer requirements.
Neither deployment option is currently offered as an available or production-ready product.
DebugScout aims to link investigation findings to source logs, relevant code locations, configurations and software revisions. AI-generated hypotheses require engineering verification; they are not confirmed root causes.
The public demo compares preloaded synthetic logs, shows precomputed evidence-linked findings, and lets an engineer approve or reject a simulated summary. It does not upload customer files, access repositories or run AI analysis. Source-aware investigation and the security and deployment capabilities described here are planned, not demonstrated production features.
Have specific security or deployment requirements? Let's discuss your engineering environment.
Discuss Security & Pilot RequirementsContact
We are looking for integration and support teams to shape the next workflows with us.
Tell us about your team, the ECUs you support and the kind of logs you work with. We will reply by email to discuss whether a pilot is a fit.
Register your interest in trying DebugScout as new workflows become available. For now, requests are handled by email; no account is created.
Or write to us directly:enquiries@debugscout.de