Learn Ml
39Development with Agentic tools like ADK, MCP, A2A for creating agentic workflows.
Getting Started
README
Learning AI/ML Development
Example - 1
A toy project implementing A2A Protocol with Google Agent Development kit on local network

Example - 2
A toy project implementing MCP (stdio), A2A Protocol with ADK

Example - 3
Implementing MCP over HTTP transport in previus project

Scripts
Scripts to run different sections of the project For all the commands below I recommend using different terminals, and start from the root directory.
Bun server for database endpoints
Bun server runs on http://0.0.0.0:3000
cd server
bun dev
Database Agent - CRUD over HTTP
DB Agent is exposed on http://0.0.0.0:8001 via A2A
cd learn_adk
uvicorn db_agent.agent:a2a_app --host 0.0.0.0 --port 8001
Chart Agent - Creates line chart for 2D Data
Chart Agent is exposed on http://0.0.0.0:8002 via A2A
cd learn_adk
uvicorn chart_agent.agent:a2a_app --host 0.0.0.0 --port 8002
File Creation Tool - MCP HTTP Server
File creation MCP server is exposed on http://0.0.0.0:8003 via HTTP transport
cd mcp_server
uv run adk_mcp_http.py
Web Scraper tool STDIO*
Web Scraping MCP server based on standard input output transport !Not required to run on any port since it runs over STDIO
Orchestrator Agent
Agent to orchestrate all the other agents
cd learn_adk
adk run orchestrator
Test agents on web using adk cli, defaults at port 8000
cd learn_adk
adk web
Crafted with ♥️ by Ayush Singh