MCP Autonomous Analyst

Autonomous Analyst

🧠 Overview

Autonomous Analyst is a local, agentic AI pipeline that:

Analyzes tabular data
Detects anomalies with Mahalanobis distance
Uses a local LLM (llama3.2:1b via Ollama) to generate interpretive summaries
Logs results to ChromaDB for semantic recall
Is fully orchestrated via the Model Context Protocol (MCP)

⚙️ Features

Component	Description
FastAPI Web UI	Friendly dashboard for synthetic or uploaded datasets
MCP Tool Orchestration	Each process step is exposed as a callable MCP tool
Anomaly Detection	Mahalanobis Distance-based outlier detection
Visual Output	Saved scatter plot of inliers vs. outliers
Local LLM Summarization	Insights generated using `llama3.2:1b` via Ollama
Vector Store Logging	Summaries are stored in ChromaDB for persistent memory
Agentic Planning Tool	A dedicated LLM tool (`autonomous_plan`) determines next steps based on dataset context
Agentic Flow	LLM + memory + tool use + automatic reasoning + context awareness

🧪 Tools Defined (via MCP)

Tool Name	Description	LLM Used
`generate_data`	Create synthetic tabular data (Gaussian + categorical)	❌
`analyze_outliers`	Label rows using Mahalanobis distance	❌
`plot_results`	Save a plot visualizing inliers vs outliers	❌
`summarize_results`	Interpret and explain outlier distribution using `llama3.2:1b`	✅
`summarize_data_stats`	Describe dataset trends using `llama3.2:1b`	✅
`log_results_to_vector_store`	Store summaries to ChromaDB for future reference	❌
`search_logs`	Retrieve relevant past sessions using vector search (optional LLM use)	⚠️
`autonomous_plan`	Run the full pipeline, use LLM to recommend next actions automatically	✅

🤖 Agentic Capabilities

Autonomy: LLM-guided execution path selection with autonomous_plan
Tool Use: Dynamically invokes registered MCP tools via LLM inference
Reasoning: Generates technical insights from dataset conditions and outlier analysis
Memory: Persists and recalls knowledge using ChromaDB vector search
LLM: Powered by Ollama with llama3.2:1b (temperature = 0.1, deterministic)

🚀 Getting Started

1. Clone and Set Up

git clone https://github.com/MadMando/mcp-autonomous-analyst.git
cd mcp-autonomous-analyst
conda create -n mcp-agentic python=3.11 -y
conda activate mcp-agentic
pip install uv
uv pip install -r requirements.txt

2. Start the MCP Server

mcp run server.py --transport streamable-http

3. Start the Web Dashboard

uvicorn web:app --reload --port 8001

Then visit: http://localhost:8000

🌐 Dashboard Flow

Step 1: Upload your own dataset or click Generate Synthetic Data
Step 2: The system runs anomaly detection on feature_1 vs feature_2
Step 3: Visual plot of outliers is generated
Step 4: Summaries are created via LLM
Step 5: Results are optionally logged to vector store for recall

📁 Project Layout

📦 autonomous-analyst/
├── server.py                  # MCP server
├── web.py                     # FastAPI + MCP client (frontend logic)
├── tools/
│   ├── synthetic_data.py
│   ├── outlier_detection.py
│   ├── plotter.py
│   ├── summarizer.py
│   ├── vector_store.py
├── static/                   # Saved plot
├── data/                     # Uploaded or generated dataset
├── requirements.txt
├── .gitignore
└── README.md

📚 Tech Stack

MCP SDK: mcp
LLM Inference: Ollama running llama3.2:1b
UI Server: FastAPI + Uvicorn
Memory: ChromaDB vector database
Data: pandas, matplotlib, scikit-learn

✅ .gitignore Additions

__pycache__/
*.pyc
*.pkl
.env
static/
data/

🙌 Acknowledgements

This project wouldn't be possible without the incredible work of the open-source community. Special thanks to:

Tool / Library	Purpose	Repository
🧠 Model Context Protocol (MCP)	Agentic tool orchestration & execution	modelcontextprotocol/python-sdk
💬 Ollama	Local LLM inference engine (`llama3.2:1b`)	ollama/ollama
🔍 ChromaDB	Vector database for logging and retrieval	chroma-core/chroma
🌐 FastAPI	Interactive, fast web interface	tiangolo/fastapi
⚡ Uvicorn	ASGI server powering the FastAPI backend	encode/uvicorn

💡 If you use this project, please consider starring or contributing to the upstream tools that make it possible.

This repo was created with the assistance of a local rag-llm using llama3.2:1b

This server cannot be installed

security - not tested

license - not found

quality - not tested

local-only server

The server can only run on the client's local machine because it depends on local resources.

A local, agentic AI pipeline that analyzes tabular data, detects anomalies, and generates interpretive summaries using local LLMs orchestrated via the Model Context Protocol.

Related MCP Servers

Model Context Provider (MCP) Server
Mark850409
-
security
F
license
-
quality
Facilitates enhanced interaction with large language models (LLMs) by providing intelligent context management, tool integration, and multi-provider AI model coordination for efficient AI-driven workflows.
Last updated -
Python
ServiceNow MCP Server
michaelbuckner
-
security
A
license
-
quality
A Model Context Protocol server enabling AI agents to access and manipulate ServiceNow data through natural language interactions, allowing users to search for records, update them, and manage scripts.
Last updated -
9
Python
MIT License
MCP Server for langfuse
z9905080
-
security
A
license
-
quality
An MCP server implementation that integrates AI assistants with Langfuse workspaces, allowing models to query LLM metrics by time range.
Last updated -
9
JavaScript
Apache 2.0
SurrealDB MCP Server
nsxdavid
-
security
-
license
-
quality
Connect your AI models to your favorite SurrealDB database, and let the LLMs do all your work for you.
Last updated -
3
JavaScript

View all related MCP servers