AI Learning MCP
No description available
Ask AI about AI Learning MCP
Powered by Claude · Grounded in docs
I know everything about AI Learning MCP. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
AI_Learning_MCP
lab 1
Context
Two Mcp tools are used. One tool is to use playwright and node to extract content from browsers. Another tools is to use a file system and node to write responses from a LLM model gpt-4.1-mini.
The screenshots below show results.
They could be found in the file https://github.com/ruihukuang/AI_Learning_MCP/blob/main/sandbox/banoffee.md.
The trace below shows the process in this workflow.
lab 2
Context
Use account.py to create a mcp server. This server is used to integrate with a LLM model gpt-4.1-mini to manage an account for a client, and answer questions about the account.
The screenshot below shows some outputs for a request.
lab 3
Context
3 Mcp servers are used. They work with LLM models to generate outputs.
The first one is a knowledge-graph based memory MCP server related to libsql. This mcp server persistently stores entities, observations about them, and relationships between them.
This mcp server helps to keep memory of entity info in the file https://github.com/ruihukuang/AI_Learning_MCP/blob/main/memory/Janice.db. It enables to continue conversations based on previous interactions.
The screenshot below show outputs for the first mcp server usages.
The second one is a web search MCP server related to Brave Search. This Mcp server is used to search info in web similar to google web searches.
The screenshot below show outputs for the second mcp server usages.
The third one is a financial data MCP server related to Polygon. A market.py file is used to interact Polygon API calls to get different share prices. A MCP server is created based on this file.
The screenshot below show outputs for the third mcp server usages.
lab 4
Context
Use a fetch mcp server and a brave mcp search server to do market research for a research agent. The agent is converted into a tool. This tool is used by a trade agent. There are 3 self created mcp servers and a polygon mcp server for this trade agent. These mcp servers are used to make API calls to get info from Polygon platform, check stock prices, buy and share stock prices, save memory of companies, research and thinking , and send transactions results to my phone using a Pushover app etc.
The final transaction outputs are shown in the screenshot below.