How infosys/infosys-agentic-foundry works
Infosys Agentic Foundry is an enterprise platform for designing, governing, and running AI agents. A user configures agents, tools, knowledge, or workflows in the studio, then submits a chat or workflow request; backend orchestration gathers context, invokes reasoning and tools, and returns results. Separate platform capabilities include evaluation, feedback learning, access control, and model-provider routing. The workflow and product branches reflect README documentation and the repository tree; sampled excerpts are partial, so uncertain internal call wiring is not shown.
Main entry points: app_container.py Platform API · centralized_agent_inference.py Agent Orchestration · App.js Foundry Studio · AskAssistant.jsx Chat Experience
Components 25
Experience & Design
- Foundry Studio
Infosys-Agentic-Foundry-Frontend/src/App.js - Chat Experience
Infosys-Agentic-Foundry-Frontend/src/components/AskAssistant/AskAssistant.jsx - Workflow Builder
Infosys-Agentic-Foundry-Frontend/src/components/Workflow/WorkflowBuilder.jsx
Platform Management
- Platform API
Infosys-Agentic-Foundry-Backend/src/api/app_container.py - Authentication & RBAC
Infosys-Agentic-Foundry-Backend/src/auth/auth_service.py - Agent Configuration
Infosys-Agentic-Foundry-Backend/src/api/agent_endpoints.py
Agent Runtime
- Agent Templates
Infosys-Agentic-Foundry-Backend/src/agent_templates - Agent Orchestration
Infosys-Agentic-Foundry-Backend/src/inference/centralized_agent_inference.py - Workflow Execution
Infosys-Agentic-Foundry-Backend/src/inference/workflow_inference.py - Inference Engines
Infosys-Agentic-Foundry-Backend/src/inference - Tool Execution
Infosys-Agentic-Foundry-Backend/src/tools - Model Services
Infosys-Agentic-Foundry-Backend/src/models/model_service.py - LiteLLM Gateway
IAF-Litellm-Server/server.py
Context & Data
- Knowledge & Memory
Infosys-Agentic-Foundry-Backend/src/utils/knowledgebase.py - Knowledge Base Service
Infosys-Agentic-Foundry-Backend/knowledgebase_server/main.py - Vector Store
Infosys-Agentic-Foundry-Backend/src/utils/postgres_vector_store_jsonb.py - Platform Data
Infosys-Agentic-Foundry-Backend/src/database/database_manager.py - Cloud File Storage
Infosys-Agentic-Foundry-Backend/src/storage/base.py
Evaluation & Operations
- Evaluation Framework
Infosys-Agentic-Foundry-Backend/src/api/evaluation_endpoints.py - Feedback Learning
Infosys-Agentic-Foundry-Backend/src/api/feedback_learning_endpoints.py - Kafka Agent Worker
Infosys-Agentic-Foundry-Backend/agent_worker/kafka_agent_worker.py - Kafka
Infosys-Agentic-Foundry-Backend/src/database/kafka_handler.py
Other components
- Platform User
- LLM Providers
- Cloud Storage Services
Connections 29
- Platform User → Foundry Studio: designs agents
- Platform User → Chat Experience: submits requests
- Foundry Studio → Platform API: configures platform
- Chat Experience → Platform API: sends chat
- Workflow Builder → Platform API: manages workflows
- Platform API → Authentication & RBAC: checks access
- Platform API → Agent Configuration: manages agents
- Agent Configuration → Agent Templates: selects template
- Platform API → Agent Orchestration: dispatches inference
- Platform API → Workflow Execution: dispatches workflow
- Workflow Execution → Agent Orchestration: runs agents
- Agent Orchestration → Knowledge & Memory: queries context
- Agent Orchestration → Inference Engines: invokes reasoning
- Agent Orchestration → Tool Execution: executes tools
- Inference Engines → Model Services: requests completion
- Model Services → LiteLLM Gateway: routes model call
- LiteLLM Gateway → LLM Providers: calls providers
- Knowledge Base Service → Vector Store: stores embeddings
- Knowledge & Memory → Vector Store: retrieves knowledge
- Platform API → Knowledge Base Service: manages knowledge
- Platform API → Platform Data: persists platform data
- Platform API → Cloud File Storage: stores uploads
- Cloud File Storage → Cloud Storage Services: reads and writes
- Platform API → Evaluation Framework: runs evaluations
- Platform API → Feedback Learning: records feedback
- Feedback Learning → Platform Data: stores learning data
- Platform API → Kafka: publishes work
- Kafka → Kafka Agent Worker: delivers messages
- Kafka Agent Worker → Agent Orchestration: runs agent jobs