How jyothish9988/ai-workflow-saas works
AWF is a self-hosted visual workflow automation product. Users configure connections, compose and save node graphs, execute them, and inspect run history; workflows can also run on cron schedules. The backend exposes authenticated APIs, persists workflow and execution state, and dispatches graph execution to node handlers and LLM services. The graph includes documented capabilities whose exact implementation paths or call sites were not sampled; those relationships are identified as README-supported rather than implying source inspection.
Main entry points: api.js REST client · execution_engine.py Execution engine · workflow.py Workflow API · WorkflowEditor.jsx Workflow editor
Components 28
Web application
- App routes
frontend/src/App.jsx - Workflow editor
frontend/src/pages/workflow/WorkflowEditor.jsx - Node registry
frontend/src/pages/workflow/nodeRegistry.js - Run and schedule views
frontend/src/pages/Executions.jsx - Connection settings
frontend/src/pages/settings/AppIntegration.jsx - REST client
frontend/src/utils/api.js
Backend APIs
- Auth API
backend/app/api/auth.py - Workflow API
backend/app/api/workflow.py - Integration API
backend/app/api/app_integrations.py - LLM connections API
backend/app/api/llm_connections.py - Schedule API
backend/app/api/schedule.py
Workflow runtime
- Execution engine
backend/app/services/execution_engine.py - Graph control flow
backend/app/services/execution/control_flow.py - Node handlers
backend/app/services/node_handlers.py - Gmail nodes
backend/app/services/gmail_nodes.py - Background run manager
backend/app/services/run_manager.py - Schedule service
backend/app/services/schedule_service.py
Connections and integrations
- LLM service
backend/app/services/llm_service.py - Ollama service
backend/app/services/ollama_service.py - Authentication and encryption
backend/app/core/security.py
Persistence
- Database session layer
backend/app/database/db.py - Workflow and run records
backend/app/models/workflow_execution.py
Other components
- User
- Browser client
- PostgreSQL
- Hosted LLM providers
- External app APIs
- Ollama server
Connections 33
- User → Browser client: uses
README.md - Browser client → App routes: opens
frontend/src/App.jsx - App routes → Workflow editor: routes to
frontend/src/App.jsx - User → Workflow editor: builds workflows
README.md - Workflow editor → Node registry: looks up nodes
frontend/src/pages/workflow/WorkflowEditor.jsx - Workflow editor → REST client: saves and runs
frontend/src/pages/workflow/WorkflowEditor.jsx - Run and schedule views → REST client: requests run data
frontend/src/utils/api.js - Connection settings → REST client: manages connections
frontend/src/utils/api.js - REST client → Auth API: calls
frontend/src/utils/api.js - REST client → Workflow API: calls
frontend/src/utils/api.js - REST client → Integration API: calls
frontend/src/utils/api.js - REST client → LLM connections API: calls
frontend/src/utils/api.js - REST client → Schedule API: calls
frontend/src/utils/api.js - Workflow API → Authentication and encryption: authenticates
backend/app/api/workflow.py - Auth API → Authentication and encryption: issues tokens
backend/app/api/auth.py - Integration API → Authentication and encryption: encrypts secrets
backend/app/api/app_integrations.py - LLM connections API → Authentication and encryption: authenticates
backend/app/api/llm_connections.py - Workflow API → Background run manager: starts runs
backend/app/api/workflow.py - Background run manager → Execution engine: executes workflow
backend/app/services/run_manager.py - Execution engine → Graph control flow: dispatches branches
backend/app/services/execution_engine.py - Execution engine → Node handlers: runs leaf nodes
backend/app/services/execution_engine.py - Node handlers → Gmail nodes: delegates Gmail
backend/app/services/node_handlers.py - Execution engine → Workflow and run records: records executions
backend/app/services/execution_engine.py - Workflow API → Workflow and run records: manages workflows and runs
backend/app/api/workflow.py - Schedule API → Schedule service: manages schedules
- Schedule service → Workflow and run records: persists schedule state
- Execution engine → LLM service: uses AI Agent
README.md - LLM service → Hosted LLM providers: sends prompts
README.md - LLM service → Ollama service: delegates local inference
backend/app/services/llm_service.py - Ollama service → Ollama server: requests inference
README.md - Node handlers → External app APIs: uses app nodes
README.md - Database session layer → PostgreSQL: stores records
README.md - Workflow and run records → Database session layer: uses ORM base
backend/app/models/workflow_execution.py