In this video, I demonstrate mainly how on mobile device you can build and debug an AI-powered web scraping pipeline using FlowOS - a distributed Petri net execution engine.
🔧 What we build:
A scraping pipeline that takes URLs, deduplicates them against a registry, sends unique URLs to Claude for content extraction, and routes results based on confidence scores.
🤖 Working with Claude Code:
Watch as Claude helps debug a stuck transition in real-time - identifying issues with host configuration, optional presets, and the correct API endpoints to manually fire transitions.
⚡ Key concepts covered:
• Petri net places as distributed token queues
• Transition inscriptions with ArcQL queries
• MAP actions for token transformation
• Emit rules for conditional routing
• Runtime vs visual PNML storage
• Master-managed vs executor-managed transitions
• The fireOnce endpoint for debugging
🛠️ Tech Stack:
• FlowOS (flow-net-master, flow-net-node, flow-net-executor)
• Claude Code CLI for pair programming
• Spring Boot reactive services
• Petri net formalism for workflow orchestration
The pipeline processes a URL through deduplication → Claude extraction → quality assessment → confidence-based routing.
#PetriNets #AI #Claude #WorkflowAutomation #FlowOS #WebScraping #Orchestration
---------------------------------------------
Agentic-Nets is a governed runtime for autonomous AI agents. Agents live inside formal Petri nets that define what they can see, what they can do, and how results flow, with typed tokens, event sourcing, a permission model, and a secrets vault.
#AgenticNets #AgenticNetOS #AIAgents #MultiAgent #PetriNets #Claude #ClaudeCode #Automation #AgentOrchestration