JK / Loading system
Intelligence in motion / System study
An AI engineering platform that helps teams understand, review and improve software systems.
SEVEN PRODUCT SURFACES / V1
RontgenAI is not presented as a future concept. Its public preview exposes seven focused engineering products, authenticated application routes and structured demonstration output.
RontgenAI gives different engineering questions their own focused product surface while keeping the output structured and inspectable.
Architecture, repository behavior, pull requests, incidents and delivery pipelines are usually investigated in separate tools. That fragmentation makes it harder to form one reliable picture of how a software system works and where engineering attention is needed.
RontgenAI is a public engineering-intelligence suite with focused workflows for architecture review, repository explanation, pull-request review, issue planning, incident analysis, CI/CD optimization and data exploration.
The flagship story follows a sample repository from connection to an engineer-ready explanation of services, dependencies, risks and next actions.
Start with a repository, architecture diagram, pull request, incident trace, dataset or CI workflow.
The selected product organizes the input into boundaries, relationships and evidence that can be inspected.
The result calls out relevant paths, architectural decisions, bottlenecks or failure signals.
Structured output makes remediation, review comments or implementation steps explicit.
Rontgen supports judgment with evidence; it does not hide the decision behind a score.
Each public v1 surface is narrow enough to understand and useful without requiring one general-purpose assistant.
Review architecture diagrams for bottlenecks, single points of failure and improvement paths.
Explore spreadsheets and SQL data through plain-language questions with inspectable query output.
Map unfamiliar repositories, surface architecture decisions and answer questions about the codebase.
Review pull requests for bugs, security concerns and potential regressions in context.
Turn an engineering issue into an implementation plan and a proposed pull request.
Correlate production evidence to support root-cause analysis and remediation.
Inspect CI workflows for critical paths, cache misses, flaky tests and duplicated work.
Public inputs enter a focused workflow, gain shared context and return as structured evidence for a human decision.
Repositories, changes, issues, incidents and delivery context provide the working material.
A shared representation connects code, architecture and workflow relationships.
Each product workflow applies its own reasoning while using the shared context.
Authentication, billing, usage and infrastructure concerns remain shared across products.
Explanations and review findings support—not replace—engineering judgment.
RontgenAI is the portfolio's flagship proof: a public suite with seven v1 product entry points and a shared visual and interaction system.
Seven focused product routes separate architecture, data, repository, review, planning, incident and CI concerns.
Each workflow constrains its input and output around one engineering task instead of presenting a generic chat surface.
Product surfaces accept diagrams, tabular data, repositories, changes, operational evidence and workflow files.
Outputs are structured and inspectable; deeper retry, idempotency and recovery evidence remains part of ongoing product hardening.
Product workflows sit behind an authenticated application boundary, while public pages expose only the product narrative.
Rontgen favors seven narrow products over one broad assistant so each workflow can express its own context and result.
The public preview is available at rontgenai.dev with seven product routes labeled v1 and live.
These captures show the public product entry, the shared workspace and focused tools for architecture, repositories, reviews, incidents and delivery.
The public experience introduces one engineering-intelligence suite rather than seven disconnected utilities.
A common dashboard keeps the seven product surfaces legible while preserving their separate jobs.
Blueprint provides a focused starting point for examining diagrams, bottlenecks and system boundaries.
Atlas concentrates repository context, mapping and explanation inside a dedicated engineering surface.
The pull-request surface makes review scope and supporting repository context explicit before analysis.
Incident material is organized around investigation and remediation instead of a generic chat thread.
Relay frames CI analysis around critical paths, repeated work and concrete optimization evidence.
These studies mirror the scanner, review and output patterns visible across the current public preview.
A navigable system map that reveals services, dependencies and review context layer by layer.
Interaction study informed by the public previewA focused review surface connecting a proposed change to architecture and repository context.
Interaction study informed by the public previewShared system understanding reduces duplicated foundations while allowing each workflow to evolve independently.
Architecture relationships and review context should be visible rather than hidden behind a single score.
Identity, billing, usage and infrastructure are treated as common services instead of being rebuilt per workflow.
RontgenAI is available as a public preview at rontgenai.dev. Seven focused v1 product surfaces are live, while four additional workflows remain on the roadmap.
Next focusDeepening the live workflows, making their evidence easier to inspect and validating them with real engineering teams.