Tool
visualpy
I got tired of explaining “what the scripts do” to people who shouldn't have to read Python to know what's running on their servers. So I built this. Point it at any Python folder. Get a visual breakdown of what every script does, which services it touches, and how files connect.
The problem.
When AI agents or developers build Python automations, non-technical stakeholders can't see what was built. They can't verify it. They can't explain it to others. They can't audit what external services it touches.
Existing tools either require rewrites (n8n, Zapier), only work for developers (pyflowchart, code2flow), need execution (LangSmith), or produce text-only output (LLM summarizers).
visualpy fills the gap: parse any existing Python folder into a non-technical visual breakdown. No execution, no config, no rewrites.
How it works.
Static analysis, no execution
Reads code as a syntax tree (Python’s AST module). Extracts meaning without running anything. A 100-script project analyzes in under 2 seconds. Safe even for code that talks to real APIs and databases.
Service mapping
Translates imports into human language. boto3 becomes “AWS S3.” slack_sdk becomes “Slack.” 45+ libraries mapped. Extensible for your own.
Step classification
Every operation is classified: API call, file I/O, database operation, data transform, decision, output. Each type gets a visual shape. A non-technical person can scan the diagram and follow the flow.
Cross-file connections
Detects import relationships, file I/O chains, and shared environment variables. Shows the dependency graph: if you change this file, these four scripts are affected.
From panic to clarity.
50 scripts
Nobody knows what they do. Good luck.
visualpy serve ./AST decode
No execution. 2 seconds.
Clarity
Scripts, services, secrets, connections. For humans.
Screenshot of a real visualpy analysis — coming soon.
Who it's for.
Operations teams
50 inherited Python scripts running in cron. Nobody knows what they do. If one breaks, nobody can debug it.
Point visualpy at the folder. See every script, every service it touches, every connection between files. In seconds.
Managers & stakeholders
A developer or AI agent built automations. You can’t read code. You can’t verify what was built.
Visual breakdown: this script fetches from Salesforce, enriches with LinkedIn, uploads results, emails the team. Is that what you asked for?
Compliance & audit
Which automations touch PII? Where does data go after processing? The auditor needs answers.
Service inventory, secret inventory, data flow tracing. Answer the compliance question visually.
Under the hood
FastAPI + HTMX + Alpine.js. Mermaid.js for diagram rendering. Zero config — just visualpy serve /path/to/scripts. Self-hosted, open source (MIT), respects existing code as sacred. One bad file doesn't kill the analysis — it gets skipped with a warning, and the rest proceeds. Because real codebases are messy. That's fine. We work with messy.
Want to make your automations visible? Or just curious? Either works.
hi@lexi-energy.com →