CogniQuery

AI-powered data analysis via Slack. Query databases, generate charts, create reports.

Installation

uv pip install -r requirements.txt

Install weasyprint for PDF generation.

Configuration

Set environment variables in .env:

SLACK_BOT_TOKEN=xoxb-your-bot-token
SLACK_APP_TOKEN=xapp-your-app-token
GEMINI_API_KEY=your-gemini-key
DATABASE_URL=postgresql://user:pass@host:port/db

Usage

Start MCP Server

# HTTP mode
python server.py http --host 127.0.0.1 --port 8010

Start Slack Bot

python slack_bot.py

Query Data

Mention bot in Slack: @CogniQuery What are top sales by region?

Bot analyzes database, creates charts, posts results.

Features

• PostgreSQL database queries
• Automatic chart generation (matplotlib)
• PDF report creation
• Slack message/file uploads
• Schema exploration
• SQL execution
• Code interpretation for advanced analysis

Requirements

• Python 3.8+
• PostgreSQL database
• Slack workspace with bot
• Gemini API key
• MCP-compatible tools

Architecture

• MCP Server: Provides Slack integration tools
• LangChain Agent: Uses Gemini LLM for analysis
• Slack Bot: Handles mentions and responses
• Tools: Schema explorer, SQL executor, code interpreter, PDF generator