How priyadarsh912/automation-ai-agent works
PersonaAI is an autonomous, persona-driven content agent with a web dashboard. Users configure an agent, trigger or await a cycle, and review published posts, editorial decisions, memory, and analytics. The sampled TypeScript server dispatches dashboard requests to the TypeScript agent runtime; the separate Python agent tree also describes an autonomous discover–judge–write workflow. Several modules and route implementations were not sampled, so those capabilities are represented without speculative internal wiring.
Main entry points: useAgent.ts Agent Controls · main.py Python Orchestrator · server.ts Agent API · index.tsx Landing Page
Components 19
Dashboard
- Landing Page
website/src/routes/index.tsx - Agent Controls
website/src/hooks/useAgent.ts - Post Feed
website/src/routes/feed.tsx - Post Detail
website/src/routes/post.$id.tsx - Analytics
website/src/routes/analytics.tsx - Memory View
website/src/routes/memory.tsx - Agent Settings
website/src/routes/settings.tsx
Web Runtime
- Agent API
website/src/server.ts - Agent Cycle
website/src/lib/agent.ts - Persisted State
website/src/lib/agent.ts
Python Agent
- Python Orchestrator
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/main.py - Topic Discovery
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/agent/discovery.py - Editorial Judge
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/agent/judge.py - Persona Writer
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/agent/persona.py - Agent Memory
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/agent/memory.py - LLM Adapter
workspace-019fe02f-8906-7b00-a122-09a1ded60640/app/agent/llm_adapter.py
External Services
- arXiv
- Gemini API
Other components
- Agent Operator
Connections 22
- Agent Operator → Landing Page: visits
- Agent Operator → Post Feed: reviews
- Agent Operator → Analytics: reviews
- Agent Operator → Memory View: reviews
- Agent Operator → Agent Settings: configures
- Landing Page → Agent Controls: launches
- Post Feed → Agent Controls: loads state
- Post Detail → Agent Controls: loads state
- Analytics → Agent Controls: loads state
- Memory View → Agent Controls: loads state
- Agent Settings → Agent Controls: updates settings
- Agent Controls → Agent API: sends requests
- Agent API → Agent Cycle: runs cycle
- Agent API → Persisted State: reads/writes
- Agent Cycle → Persisted State: reads/writes
- Python Orchestrator → Topic Discovery: requests topics
- Python Orchestrator → Editorial Judge: evaluates topics
- Python Orchestrator → Persona Writer: builds post
- Python Orchestrator → LLM Adapter: generates text
- Python Orchestrator → Agent Memory: uses memory
- Topic Discovery → arXiv: fetches papers
- LLM Adapter → Gemini API: generates content