How vamsi-krishna-b/ai-agents works
This is an autonomous YouTube channel manager. A user starts or monitors a job in the web UI; the API records it and dispatches background work, then the orchestrator runs research, scripting, media, voice, editing, thumbnail, and upload stages. The product also exposes approvals, scheduling, settings, analytics, and live logs. Source excerpts establish the main dispatch and orchestration links; some unsampled page and agent details are represented only where the README or source index supports them.
Main entry points: orchestrator.py Pipeline orchestrator · client.ts HTTP API client · App.tsx Web app routes · main.py FastAPI and scheduler
Components 31
Web console
- Web app routes
frontend/src/App.tsx - Dashboard
frontend/src/pages/Dashboard.tsx - Pipeline monitor
frontend/src/pages/PipelineMonitor.tsx - Script editor
frontend/src/pages/ScriptEditor.tsx - Thumbnail page
frontend/src/pages/Thumbnail.tsx - Calendar
frontend/src/pages/Calendar.tsx - Analytics page
frontend/src/pages/Analytics.tsx - Settings page
frontend/src/pages/Settings.tsx - HTTP API client
frontend/src/api/client.ts - Live updates
frontend/src/hooks/useWebSocket.ts
API and jobs
- FastAPI and scheduler
backend/main.py - Job routes
backend/api/jobs.py - Settings routes
backend/api/settings.py - Analytics routes
backend/api/analytics.py - Celery pipeline tasks
backend/tasks/pipeline_tasks.py
Content pipeline
- Pipeline orchestrator
backend/agents/orchestrator.py - Topic research
backend/agents/research_agent.py - Script generation
backend/agents/script_agent.py - Video clips
backend/agents/media_agent.py - Voiceover
backend/agents/voice_agent.py - Video editing
backend/agents/editor_agent.py - Thumbnail generation
backend/agents/thumbnail_agent.py - YouTube upload
backend/agents/upload_agent.py
AI and publishing
- Provider-aware text
backend/utils/llm.py - Gemini services
backend/utils/gemini.py - Veo generation
backend/utils/veo.py - YouTube API
backend/utils/youtube.py
Persistence
- MongoDB collections
backend/db/mongo.py
Other components
- Channel manager
- YouTube
- Gemini / Groq / Veo
Connections 48
- Channel manager → Web app routes: uses
frontend/src/App.tsx - Web app routes → Dashboard: routes to
frontend/src/App.tsx - Web app routes → Pipeline monitor: routes to
frontend/src/App.tsx - Web app routes → Script editor: routes to
frontend/src/App.tsx - Web app routes → Thumbnail page: routes to
frontend/src/App.tsx - Web app routes → Calendar: routes to
frontend/src/App.tsx - Web app routes → Analytics page: routes to
frontend/src/App.tsx - Web app routes → Settings page: routes to
frontend/src/App.tsx - Dashboard → HTTP API client: calls
frontend/src/pages/Dashboard.tsx - Pipeline monitor → HTTP API client: calls
frontend/src/pages/PipelineMonitor.tsx - Pipeline monitor → Live updates: subscribes
frontend/src/pages/PipelineMonitor.tsx - Pipeline monitor → FastAPI and scheduler: receives updates
- Script editor → HTTP API client: calls
frontend/src/pages/ScriptEditor.tsx - Thumbnail page → HTTP API client: calls
frontend/src/pages/Thumbnail.tsx - Calendar → HTTP API client: calls
frontend/src/pages/Calendar.tsx - Analytics page → HTTP API client: calls
frontend/src/pages/Analytics.tsx - Settings page → HTTP API client: calls
frontend/src/pages/Settings.tsx - HTTP API client → Job routes: requests jobs
frontend/src/api/client.ts - HTTP API client → Settings routes: requests settings
frontend/src/api/client.ts - HTTP API client → Analytics routes: requests analytics
frontend/src/api/client.ts - FastAPI and scheduler → Job routes: mounts router
backend/main.py - FastAPI and scheduler → Settings routes: mounts router
backend/main.py - FastAPI and scheduler → Analytics routes: mounts router
backend/main.py - Job routes → Pipeline orchestrator: creates jobs
backend/api/jobs.py - Job routes → Celery pipeline tasks: dispatches
backend/api/jobs.py - Job routes → MongoDB collections: reads and writes
backend/api/jobs.py - Settings routes → MongoDB collections: reads and writes
backend/api/settings.py - Analytics routes → MongoDB collections: reads and writes
backend/api/analytics.py - Analytics routes → YouTube API: fetches metrics
backend/api/analytics.py - Celery pipeline tasks → Pipeline orchestrator: runs pipeline
backend/tasks/pipeline_tasks.py - Celery pipeline tasks → MongoDB collections: connects and closes
backend/tasks/pipeline_tasks.py - FastAPI and scheduler → Pipeline orchestrator: schedules jobs
backend/main.py - FastAPI and scheduler → Celery pipeline tasks: dispatches scheduled
backend/main.py - FastAPI and scheduler → MongoDB collections: manages connection
backend/main.py - Pipeline orchestrator → MongoDB collections: stores job state
backend/agents/orchestrator.py - Pipeline orchestrator → Topic research: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → Script generation: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → Video clips: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → Voiceover: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → Video editing: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → Thumbnail generation: runs stage
backend/agents/orchestrator.py - Pipeline orchestrator → YouTube upload: runs stage
backend/agents/orchestrator.py - Topic research → Provider-aware text: generates topics
backend/agents/research_agent.py - Script generation → Provider-aware text: generates script
backend/agents/script_agent.py - Video clips → Veo generation: generates clips
backend/agents/media_agent.py - Voiceover → Gemini services: generates audio
backend/agents/voice_agent.py - Thumbnail generation → Gemini services: generates image
backend/agents/thumbnail_agent.py - YouTube upload → YouTube API: uploads video
backend/agents/upload_agent.py