AssistantAI_MCP

Zahaawii/AssistantAI_MCP

3.2

If you are the rightful owner of AssistantAI_MCP and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

This project implements a Model Context Protocol (MCP) server and a static MCP client, facilitating seamless communication and tool execution.

MCP Server Project

This project implements a Model Context Protocol (MCP) server.

The system is now fully integrated with the AssistantAI and provides dynamic tool execution, data management, authentication, CI/CD, and deployment.
It has evolved into a feature-rich backend agent supporting real workflows across ChromaDB and your blog platform.


🚀 Current Capabilities

🧩 MCP Server

The server fully supports interaction with ChromaDB and your Blog API.

ChromaDB Tools
  • Read documents
  • Upload documents
  • Delete documents
  • AssistantAI can autonomously create/store new knowledge
Blog Tools
  • Read all blog posts
  • Read posts by author
  • Log in and obtain a JWT token
  • Delete blog posts
  • Create structured blog posts

🤖 AssistantAI Integration

AssistantAI is fully connected and can:

  • Upload new articles into ChromaDB
  • Read, create, and delete blog posts
  • Authenticate as a user or service account
  • Perform actions as an intelligent agent

🔄 CI/CD + Testing

The project includes:

  • CI/CD pipeline using GitHub Actions
  • Unit tests (Mockito, JUnit)
  • Integration tests validating MCP → Backend → ChromaDB flow

🌍 Deployment

The MCP server is live and available for external MCP clients at:

http://mcp.zaak.dk/mcp


📚 Resources Used During Development

(And many additional references explored during implementation.)


🧭 Next Steps

1. Improve Backend Robustness & Error Handling

  • Strong typed error responses
  • Input validation
  • Fallback logic
  • Logging + tracing

2. Expand Toolset

  • Blog comment tools
  • User profile tools
  • Multi-step workflows
  • Combined ChromaDB + blog tools

3. Enhance AssistantAI Behavior

  • Better tool-calling logic
  • Planning + multi‑step reasoning
  • Dynamic context retrieval
  • More autonomous agent abilities

4. Optional Dashboard UI

  • Browse ChromaDB
  • Trigger tools manually
  • View logs, metadata, and execution traces

📄 License

MIT License.