Скачать или смотреть My LangGraph Project 3| Advanced Multi-Agent System with, Planners & Human-in-the-Loop for Saleforce

The system is designed with a supervisor agent that intelligently coordinates work between two specialized sub-agents, each with its own independent planner. The workflow includes a Human-in-the-Loop approval stage, making the planning pipeline transparent and controllable.

⭐ What This System Includes
Supervisor Agent
Routes user queries to the correct planner and sub-agent based on context.

Two Sub-Agents
Each agent is specialized for different domains and powered by advanced reasoning.

Two Independent Planners
Each planner creates structured plans depending on the query domain.

Human-in-the-Loop Gate
Approve
Edit
Reject

Re-plan
Nothing proceeds until the human confirms.

MCP (Model Context Protocol) Integration
Tools auto-loaded from MCP over SSE transport.
Sub-agents use these tools as actions inside the reasoning loop.

Deep Agent Architecture
Automatic tool discovery, planner → agent routing, conversation memory, checkpointing, and graceful error handling.

Dynamic Plan Execution
Once the plan is human-approved, the system executes it through the correct sub-agent and returns the final structured output.

🎥 What You Will See in This Demo
Full planning sequence
Supervisor’s reasoning
Planner selection logic
Human approval step
Tool-calling through MCP/SSE
Sub-agent execution
Final end-to-end result

This is a complete showcase of how advanced multi-agent systems can be built with LangGraph, LangChain, MCP, and modern agent architecture patterns.

📌 Join My DataLinkk Skool Community (FREE)
👉 https://www.skool.com/qya-automations...

📎 Useful Links
🌐 Agency – https://datalinkk.com
📺 YouTube –    / @ahmed_ai67  
💼 LinkedIn –   / sarfaraz-ahmed07  
💻 GitHub – https://github.com/Sarfaraz021