{"id":50548790,"url":"https://github.com/designer-coderajay/enterprise-agentic-ai-platform","last_synced_at":"2026-06-04T01:02:40.392Z","repository":{"id":355021624,"uuid":"1226453608","full_name":"designer-coderajay/enterprise-agentic-ai-platform","owner":"designer-coderajay","description":"Production-grade multi-agent AI platform with LangGraph v0.3, MCP servers, hybrid RAG (Qdrant + LlamaIndex), FastAPI WebSocket streaming, and Next.js real-time UI. Deployed on AWS EKS via Terraform.","archived":false,"fork":false,"pushed_at":"2026-05-01T12:42:14.000Z","size":44,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-01T14:18:37.828Z","etag":null,"topics":["agentic-ai","anthropic","fastapi","kubernetes","langchain","langgraph","llamaindex","llm","mcp","model-context-protocol","nextjs","python","qdrant","rag","terraform"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/designer-coderajay.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-01T12:20:51.000Z","updated_at":"2026-05-01T12:42:17.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/designer-coderajay/enterprise-agentic-ai-platform","commit_stats":null,"previous_names":["designer-coderajay/enterprise-agentic-ai-platform"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/designer-coderajay/enterprise-agentic-ai-platform","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/designer-coderajay%2Fenterprise-agentic-ai-platform","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/designer-coderajay%2Fenterprise-agentic-ai-platform/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/designer-coderajay%2Fenterprise-agentic-ai-platform/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/designer-coderajay%2Fenterprise-agentic-ai-platform/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/designer-coderajay","download_url":"https://codeload.github.com/designer-coderajay/enterprise-agentic-ai-platform/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/designer-coderajay%2Fenterprise-agentic-ai-platform/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33886154,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-03T02:00:06.370Z","response_time":59,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agentic-ai","anthropic","fastapi","kubernetes","langchain","langgraph","llamaindex","llm","mcp","model-context-protocol","nextjs","python","qdrant","rag","terraform"],"created_at":"2026-06-04T01:02:37.037Z","updated_at":"2026-06-04T01:02:40.387Z","avatar_url":"https://github.com/designer-coderajay.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Enterprise Agentic AI Platform\n\n\u003e **Horizontal multi-agent knowledge platform with MCP integration — production-ready on AWS EKS.**\n\u003e Built with LangGraph v0.3, FastMCP, LlamaIndex RAG, and real-time WebSocket streaming.\n\n---\n\n## What This Does\n\nA universal agentic AI backend that any enterprise can deploy to answer complex questions by:\n\n1. Retrieving relevant context from enterprise knowledge bases (RAG with Qdrant)\n2. Planning a multi-step approach using a Planner agent\n3. Executing tool calls via MCP servers (databases, documents, notifications)\n4. Critiquing the answer and revising up to 3× if quality is insufficient\n5. Streaming every agent reasoning step to the UI in real time via WebSocket\n\n**Industries**: Financial services, healthcare, legal, manufacturing, e-commerce — any domain with large internal knowledge bases.\n\n---\n\n## Architecture\n\n```\n┌─────────────────────────┐   WebSocket/REST   ┌────────────────────────────────────┐\n│     Next.js 14 UI       │ ◄────────────────► │        FastAPI Backend             │\n│   Real-time streaming   │                    │           Port 8000                │\n│   Agent trace sidebar   │                    └──────────────┬─────────────────────┘\n└─────────────────────────┘                                   │\n                                                              ▼\n                                          ┌───────────────────────────────────┐\n                                          │    LangGraph StateGraph           │\n                                          │                                   │\n                                          │  [Memory Retrieval]               │\n                                          │       ↓                           │\n                                          │  [Planner] → plan steps           │\n                                          │       ↓                           │\n                                          │  [Executor] → MCP tool calls      │\n                                          │       ↓                           │\n                                          │  [Critic] → score \u0026 critique      │\n                                          │       ↓ (if needs_revision)       │\n                                          │  [Executor] → retry (max 3×)      │\n                                          └───────────────┬───────────────────┘\n                                                          │\n                          ┌───────────────────────────────┼──────────────────────┐\n                          ▼                               ▼                      ▼\n                  ┌──────────────┐              ┌──────────────┐      ┌──────────────────┐\n                  │ Postgres MCP │              │ Document MCP │      │ Notification MCP │\n                  │   Port 8001  │              │  Port 8002   │      │   Port 8003      │\n                  │ query_db     │              │ extract_pdf  │      │ send_slack       │\n                  │ list_tables  │              │ search_doc   │      │ send_webhook     │\n                  └──────────────┘              └──────────────┘      └──────────────────┘\n                          │                               │\n                          ▼                               ▼\n                   PostgreSQL 16                     Qdrant Vector DB\n                   (asyncpg)                         (hybrid search)\n```\n\n---\n\n## Key Technical Stack\n\n| Layer | Technology |\n|-------|-----------|\n| Agent Orchestration | LangGraph v0.3 (StateGraph, checkpointing, time-travel) |\n| LLM | Claude Sonnet via LangChain |\n| MCP Servers | FastMCP + Streamable HTTP transport (OAuth 2.1 ready) |\n| RAG | LlamaIndex + Qdrant (hybrid dense+sparse search) + Cohere reranking |\n| Embeddings | OpenAI `text-embedding-3-large` |\n| API | FastAPI async + WebSocket streaming |\n| Frontend | Next.js 14 App Router + shadcn/ui + Zustand |\n| Observability | LangSmith + Langfuse + Prometheus + Grafana |\n| Infra | EKS + Terraform + GitHub Actions + ArgoCD |\n\n---\n\n## Quick Start\n\n```bash\ncd ~/Desktop/enterprise-agentic-ai-platform\ncp .env.example .env\n# Fill in: ANTHROPIC_API_KEY, LANGSMITH_API_KEY, OPENAI_API_KEY\n\n# Full stack with Docker Compose\nmake dev-full\n\n# Or individual services\nmake mcp-start    # Start 3 MCP servers (:8001, :8002, :8003)\nmake dev          # FastAPI backend on :8000\ncd frontend \u0026\u0026 npm install \u0026\u0026 npm run dev   # Next.js on :3000\n```\n\n---\n\n## MCP Servers\n\n| Server | Port | Tools |\n|--------|------|-------|\n| `postgres_mcp` | 8001 | `query_database`, `list_tables`, `insert_record` |\n| `document_mcp` | 8002 | `extract_text_from_pdf`, `extract_tables_from_pdf`, `search_document` |\n| `notification_mcp` | 8003 | `send_slack_message`, `send_webhook`, `create_slack_reminder` |\n\nAll MCP servers use **Streamable HTTP transport** (the April 2026 standard — SSE deprecated).\n\n---\n\n## Agent Loop\n\n```python\n# Planner → Executor → Critic → (revise?) loop\ndef should_revise(state: AgentState):\n    if state.needs_revision and state.revision_count \u003c MAX_REVISIONS:  # MAX=3\n        return \"executor\"\n    return \"END\"\n\n# Graph nodes\ngraph.add_node(\"memory_retrieval\", memory_retrieval_node)\ngraph.add_node(\"planner\", planner_node)\ngraph.add_node(\"executor\", executor_node)\ngraph.add_node(\"critic\", critic_node)\n\n# Critic loop — prevents infinite revision\ngraph.add_conditional_edges(\"critic\", should_revise, {\n    \"executor\": \"executor\",\n    \"END\": END\n})\n```\n\n---\n\n## RAG Pipeline\n\n- **Ingestion**: LlamaIndex `IngestionPipeline` with `SemanticSplitterNodeParser`\n- **Vector Store**: Qdrant with `enable_hybrid=True` (dense + sparse BM25)\n- **Retrieval**: Score-threshold filtering + sorted by relevance\n- **Supported Sources**: S3 URIs, local PDFs, URLs\n\n```bash\n# Ingest documents\ncurl -X POST http://localhost:8000/api/v1/documents/ingest \\\n  -F \"files=@annual_report.pdf\"\n```\n\n---\n\n## WebSocket Streaming\n\n```javascript\n// Client receives agent step-by-step reasoning in real time\nws.onmessage = (event) =\u003e {\n  const { type, node, data } = JSON.parse(event.data);\n  if (type === \"node_update\") {\n    // Show agent reasoning: planner → executor → critic\n    appendAgentStep(node, data);\n  }\n};\n```\n\n---\n\n## Running Tests\n\n```bash\nmake test       # All tests\nmake test-cov   # With coverage\npytest tests/test_agents.py -v -s   # Agent tests with output\n```\n\n---\n\n## Deployment\n\n```bash\n# Terraform: EKS (us-east-1) + RDS + ElastiCache + S3\ncd infra/terraform \u0026\u0026 terraform init \u0026\u0026 terraform apply\n\n# CI/CD: GitHub Actions\n# push to main → pytest → ECR push → kubectl rolling update\n```\n\n**Kubernetes**: HPA scales 3→20 replicas on CPU (70%) and memory (80%).\n\n---\n\n## Resume Bullet Points 🚀\n\n- **Designed** production multi-agent RAG platform using LangGraph v0.3 StateGraph with Planner-Executor-Critic loop, supporting 3× self-revision and human-in-the-loop interruption\n- **Built** 3 production MCP servers (PostgreSQL, Document, Notification) using FastMCP + Streamable HTTP transport (April 2026 standard) with asyncpg connection pooling\n- **Implemented** hybrid RAG pipeline (dense + sparse search) with LlamaIndex, Qdrant, semantic chunking, and Cohere reranking; ingests PDFs, S3 documents, URLs\n- **Architected** real-time WebSocket streaming API with FastAPI and Next.js 14 showing live agent reasoning; falls back to REST seamlessly\n- **Integrated** LangSmith + Langfuse tracing, Prometheus custom metrics (tokens, latency, tool calls), and Grafana dashboards for full observability\n- **Deployed** on AWS EKS with HPA (3→20 replicas), Terraform IaC, ArgoCD GitOps, GitHub Actions CI/CD\n\n---\n\n## Folder Structure\n\n```\nenterprise-agentic-ai-platform/\n├── backend/\n│   ├── agents/         # state, graph, planner, executor, critic, memory\n│   ├── rag/            # ingestion pipeline + retriever (LlamaIndex + Qdrant)\n│   ├── mcp_servers/    # postgres_mcp, document_mcp, notification_mcp\n│   ├── api/            # FastAPI routes (health, documents)\n│   ├── observability/  # LangSmith + Prometheus metrics\n│   └── main.py         # FastAPI app (REST + WebSocket)\n├── frontend/           # Next.js 14 chat UI + agent trace sidebar\n├── infra/\n│   ├── k8s/            # Deployment, HPA, ConfigMaps\n│   ├── terraform/      # EKS, RDS, ElastiCache, S3\n│   ├── docker/         # Backend + frontend Dockerfiles\n│   └── .github/        # CI/CD workflow\n├── tests/              # pytest-asyncio agent tests\n├── docker-compose.yml  # Full local stack (11 services)\n└── Makefile            # All dev commands\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdesigner-coderajay%2Fenterprise-agentic-ai-platform","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdesigner-coderajay%2Fenterprise-agentic-ai-platform","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdesigner-coderajay%2Fenterprise-agentic-ai-platform/lists"}