Tracking MCP
A generic MCP server for tracking any entity type (e.g., weight, workouts, books) without predefined schemas, using SQLite with JSON columns for flexible data storage and querying.
README
Tracking MCP
Generic MCP server for tracking any entity type with schema-less JSON Hybrid storage.
Track body weight, daily scorecards, fitness sessions, books, or any custom entity without defining rigid schemas. Auto-discovery, self-documenting, and SQL-queryable.
Features
- Schema-less Design: Track any entity type (weight, scorecard, fitness, books, custom) without ALTER TABLE
- Auto-Discovery: Entity types automatically registered on first use
- Self-Documenting: MCP Resources expose schema examples and usage guides
- SQL-Queryable: Use
json_extract()for advanced analytics - Local-First: Privacy-friendly, zero external dependencies
- Hybrid Storage: SQLite with JSON columns for flexibility + performance
- CRUD Operations: Insert, update, query, delete via MCP Tools
- Built-in Prompts: Pre-configured templates for common tracking scenarios
Installation
Using uvx (Recommended)
# Run directly without installation
uvx tracking-mcp
# Or install globally
pip install tracking-mcp
From Source
git clone https://github.com/mindfullabai/tracking-mcp.git
cd tracking-mcp
pip install -e .
Initialize Database
Database is auto-created on first use at the path specified in DB_PATH environment variable (defaults to ~/tracking.db).
Quick Start
Claude Desktop Configuration
Add to your Claude Desktop MCP settings (.mcp.json or ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"tracking-mcp": {
"type": "stdio",
"command": "uvx",
"args": ["tracking-mcp"],
"env": {
"DB_PATH": "/path/to/your/data/tracking.db"
}
}
}
}
Alternative (with pip install):
{
"mcpServers": {
"tracking-mcp": {
"type": "stdio",
"command": "tracking-mcp",
"env": {
"DB_PATH": "/path/to/your/data/tracking.db"
}
}
}
}
Basic Usage
From Claude Desktop, you can now:
Track my weight: 72.5kg today
Show me my weight trend for the last 30 days
Log workout: HYROX for 45 minutes today
MCP Server Specification
Tools (4)
1. track_event
Insert or update tracking event for any entity type.
Parameters:
entity_type(string, required): Entity type (e.g., 'weight', 'scorecard', 'fitness', 'book', or custom)date(string, required): Event date in YYYY-MM-DD formatdata(object, required): Entity-specific data (schema-free JSON)entity_id(string, optional): Unique ID for entity instance (e.g., 'book_atomic_habits')
Example:
track_event(
entity_type="weight",
date="2026-01-14",
data={"weight_kg": 72.8, "day_type": "MAR", "source": "manual"}
)
2. query_events
Query tracking events with filters.
Parameters:
entity_type(string, optional): Filter by entity typeentity_id(string, optional): Filter by entity IDstart_date(string, optional): Start date (inclusive)end_date(string, optional): End date (inclusive)limit(integer, optional): Maximum results (default: 100)
Example:
query_events(
entity_type="weight",
start_date="2025-12-15",
end_date="2026-01-14",
limit=30
)
3. delete_event
Delete tracking event by ID.
Parameters:
event_id(integer, required): Event ID to delete
4. list_entity_types
Get all registered entity types with schema examples.
Returns: JSON array of entity types with descriptions and schema examples.
Resources (3)
1. tracking://schema/entity_types
List of all registered entity types with schema examples (JSON).
2. tracking://docs/usage
Usage guide for tracking new entity types dynamically (Markdown).
3. tracking://stats/summary
Current statistics: total events, entity types, date range, events by type (JSON).
Prompts (3)
1. track-weight
Template for tracking body weight.
Arguments: weight_kg, date
2. track-workout
Template for logging workout session.
Arguments: workout_type, duration_min, date
3. query-trend
Get trend data for entity type over date range.
Arguments: entity_type, days (default: 30)
Database Schema
tracking_events Table
| Column | Type | Description |
|---|---|---|
id |
INTEGER PRIMARY KEY | Auto-increment ID |
entity_type |
TEXT | Entity type ('weight', 'scorecard', etc.) |
entity_id |
TEXT | Optional unique ID for entity instances |
date |
DATE | Event date (YYYY-MM-DD) |
data |
JSON | Schema-free JSON data |
created_at |
TIMESTAMP | Auto-generated creation timestamp |
updated_at |
TIMESTAMP | Auto-updated modification timestamp |
Indexes: entity_type, date, entity_id
entity_types Table
| Column | Type | Description |
|---|---|---|
entity_type |
TEXT PRIMARY KEY | Entity type name |
description |
TEXT | Human-readable description |
schema_example |
JSON | Example JSON schema |
created_at |
TIMESTAMP | Registration timestamp |
updated_at |
TIMESTAMP | Last update timestamp |
Pre-seeded entity types: weight, scorecard, fitness, book
Advanced Usage Examples
Track Custom Entity Type
# Sleep quality tracking (auto-registered)
track_event(
entity_type="sleep_quality",
date="2026-01-14",
data={
"hours": 7.5,
"quality_score": 8,
"dreams": True,
"interruptions": 2,
"notes": "Felt refreshed"
}
)
Track Entity with Unique ID
# Reading progress for specific book
track_event(
entity_type="book",
entity_id="book_atomic_habits",
date="2026-01-14",
data={
"title": "Atomic Habits",
"author": "James Clear",
"current_page": 150,
"total_pages": 320,
"rating": 5
}
)
Query with Filters
# Get all weight entries for January 2026
query_events(
entity_type="weight",
start_date="2026-01-01",
end_date="2026-01-31"
)
# Get all entries for specific book
query_events(
entity_type="book",
entity_id="book_atomic_habits"
)
Update Existing Event
To update an event, call track_event() with the same entity_type + date (+ entity_id if used). The tool will automatically UPDATE instead of INSERT.
SQL Analytics
Since data is stored in SQLite with JSON columns, you can run advanced analytics:
Weight Trend (Last 30 Days)
SELECT
date,
json_extract(data, '$.weight_kg') as weight,
json_extract(data, '$.delta_kg') as delta
FROM tracking_events
WHERE entity_type = 'weight'
AND date >= date('now', '-30 days')
ORDER BY date DESC;
Scorecard Weekly Average
SELECT
strftime('%Y-W%W', date) as week,
AVG(CAST(json_extract(data, '$.total_score') AS INTEGER)) as avg_score,
COUNT(*) as days
FROM tracking_events
WHERE entity_type = 'scorecard'
AND date >= date('now', 'weekday 0', '-7 days')
GROUP BY week;
Fitness Volume by Workout Type (This Month)
SELECT
json_extract(data, '$.workout_type') as type,
COUNT(*) as sessions,
SUM(CAST(json_extract(data, '$.duration_min') AS INTEGER)) as total_minutes,
AVG(CAST(json_extract(data, '$.duration_min') AS INTEGER)) as avg_minutes
FROM tracking_events
WHERE entity_type = 'fitness'
AND date >= date('now', 'start of month')
GROUP BY type;
Project Structure
tracking-mcp/
├── data/
│ ├── tracking.db # SQLite database
│ └── schema.sql # Database schema
├── tracking_mcp/
│ ├── tracking_server.py # MCP server implementation
│ └── __init__.py
├── tests/
│ └── test_server.py
├── pyproject.toml
├── LICENSE
├── CHANGELOG.md
└── README.md
Architecture Decisions
Why JSON Hybrid (SQLite + JSON)?
- ✅ Flexibility: Add new entity types without schema migrations
- ✅ Performance: SQLite indexes +
json_extract()for fast queries - ✅ SQL-queryable: Standard SQL for analytics
- ❌ EAV alternative: Too many JOINs, poor performance for analytics
Why Custom MCP vs Official SQLite MCP?
- ✅ Auto-discovery: New entity types registered automatically
- ✅ Self-documenting: Resources expose schemas and usage
- ✅ Dynamic: No rigid schema required
- ❌ Official SQLite MCP: Requires predefined schema
Why SQLite vs PostgreSQL?
- ✅ Zero setup: File-based, no server required
- ✅ Local-first: Privacy-friendly for personal tracking
- ✅ Sufficient: Perfect for single-user personal use
- ❌ PostgreSQL: Unnecessary overhead for personal tracking
Development
Run Tests
pytest
Code Quality
# Format code
black mcp_server/
# Lint
ruff check mcp_server/
Install Development Dependencies
pip install -e ".[dev]"
Troubleshooting
RuntimeWarning: coroutine 'main' was never awaited
If you see this error when running the server:
<coroutine object main at 0x...>
RuntimeWarning: coroutine 'main' was never awaited
This was fixed in version 1.0.1. Update to the latest version:
pip install --upgrade tracking-mcp
# or with uvx
uvx --refresh tracking-mcp
Root cause: Python CLI entry points from setuptools expect synchronous main() functions. Version 1.0.1+ includes a sync wrapper that properly handles the async MCP server.
Version History
See CHANGELOG.md for version history.
Current version: 1.0.1 (Async entry point fix)
Related Projects
- viz-mcp: Companion MCP server for auto-generating data visualizations from tracking data
- work-hub: Personal productivity system using tracking-mcp for daily scorecard and habit tracking
License
MIT License - see LICENSE file for details.
Author
Mario Mosca - GitHub
Contributing
Contributions welcome! Please open an issue or pull request.
Support
For issues, questions, or feature requests, please open an issue on GitHub: https://github.com/mariomosca/tracking-mcp/issues
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.