CS-13 — Automation & AI · Architecture & performance
AI-assisted platform dependency visualiser
Internal tool that analyses repositories and discovers and visualises the dependencies between microservices, endpoints, Lambda functions, shared libraries and infrastructure resources.
- My role
- Ander's own initiative within the OPIS project (Wizeline); design and development of the tool.
- Period
- Not documented
- Disclosure
- Restricted — high-level only
- Technologies
- LangGraphLangflowAnthropic APIClaudeTypeScriptNext.js
- Last reviewed
- 13 November 2026
Context
On enterprise platforms with many repositories, microservices, endpoints, Lambda functions, shared libraries and integrations between services, understanding the full dependency graph by hand is increasingly difficult.
Challenges
- 01Discovering dependencies spread across many repositories and resources.
- 02Keeping the map up to date as the platform changes.
Solution & architecture
An agentic application built with Langflow and LangGraph on top of the Claude API (Anthropic) that inspects code and repositories, discovers and maps dependencies and visualises the relationships between components, with full scans and recurring scan workflows to help assess the impact of changes.
Key decisions
Not documented yet
Trade-offs
Not documented yet
Results
In use by the project's engineering team.
Lessons learned
Not documented yet
Intended value: faster architecture exploration, better knowledge of dependencies, change impact analysis and support for maintenance and modernisation. The concept is described, not the confidential architecture or the client's repositories. Frontend built with the same stack as this portfolio (Next.js and TypeScript).
Ask about this case study
Questions here are answered with the case study “AI-assisted platform dependency visualiser” as context.