How modesi/publicreponistproject works
This repository is a collection of LLM client examples and related chat, agent, and MCP integrations rather than one unified application. The simplest user workflow sends text or image input to a configured model and prints its response. The richer MCP workflow uses a chat interface to obtain tool choices, call server tools, and return results to the model for a final answer. Other branches include a file-reading agent and an OpenAI-compatible API proxy. Some server implementations are unsampled, so only documented or directly evidenced relationships are shown.
Main entry points: main.py CLI Chat · app.py Chat UI · chatbot.py Chat Loop · mcp_server MCP Tools
Components 18
LLM Examples
- Text Query
example_text.py - Image Query
example_image.py
MCP Chat
- Chat UI
resources/mcp_client/app.py - Chat Loop
resources/mcp_client/chatbot.py - Tool Formatting
resources/mcp_client/llm_utils.py - CLI Chat
resources/mcp_client/main.py - MCP Client
resources/mcp_client/client.py - Client Config
resources/mcp_client/config.py
Agent Demo
- File Agent
resources/agent_demo/agent_demo.py - Sample Files
resources/agent_demo/sample_directory - MCP File Agent
resources/agent_demo/agent_demo_v2.py
API And Tools
- MCP Tools
resources/mcp_server - Ontology Server
resources/mcp_server/ontology-server.py - Calculator Server
resources/mcp_server/calculator.py - OpenAI-Compatible Proxy
rag_execution/openai.py
Other components
- User
- LLM API
- MCP Service
Connections 25
- User → Text Query: enters prompt
example_text.py - Text Query → LLM API: sends text
example_text.py - LLM API → Text Query: returns response
example_text.py - User → Image Query: asks about image
example_image.py - Image Query → LLM API: sends image query
example_image.py - LLM API → Image Query: returns response
example_image.py - User → Chat UI: enters message
resources/mcp_client/app.py - Chat UI → Chat Loop: processes turn
resources/mcp_client/app.py - Chat Loop → Tool Formatting: requests LLM response
resources/mcp_client/chatbot.py - Chat Loop → MCP Tools: calls tools
resources/mcp_client/chatbot.py - Chat UI → Client Config: loads settings
resources/mcp_client/app.py - Chat UI → LLM API: requests completion
resources/mcp_client/app.py - CLI Chat → Chat Loop: processes turns
resources/mcp_client/main.py - CLI Chat → Tool Formatting: formats tools
resources/mcp_client/main.py - CLI Chat → MCP Tools: calls tools
resources/mcp_client/main.py - CLI Chat → Client Config: loads settings
resources/mcp_client/main.py - MCP Client → Client Config: reads settings
resources/mcp_client/client.py - MCP Client → MCP Tools: calls tools
resources/mcp_client/client.py - User → File Agent: submits query
resources/agent_demo/agent_demo.py - File Agent → LLM API: runs model
resources/agent_demo/agent_demo.py - File Agent → Sample Files: lists and reads
resources/agent_demo/agent_demo.py - User → MCP File Agent: submits query
resources/agent_demo/agent_demo_v2.py - MCP File Agent → LLM API: runs model
resources/agent_demo/agent_demo_v2.py - MCP File Agent → MCP Service: connects to
resources/agent_demo/agent_demo_v2.py - MCP Service → MCP Tools: serves tools