The Fatebook MCP Server integrates with Fatebook to enable AI assistants like Claude to create, manage, and track predictions through the Model Context Protocol (MCP). Key capabilities include:
Create Questions: Make new predictions with initial forecasts (0-100% probability)
List & Search Questions: View predictions with filtering options by resolution status or search terms
Get Question Details: Retrieve comprehensive information about specific questions using their ID
Add Forecasts: Update or add new predicted probabilities on existing questions
Add Comments: Document reasoning and notes on predictions
Resolve Questions: Mark questions as resolved (YES/NO/AMBIGUOUS)
Edit Questions: Update titles, resolve dates, and notes
Delete Questions: Remove unwanted predictions
Count Forecasts: Track forecasting activity for specific users
Source code is hosted on GitHub with CI workflows and issue tracking
The MCP server has been tested and verified to work on macOS 14.5
The MCP server is distributed and installed through PyPI as the fatebook-mcp package
Used for running integration tests to verify the MCP server setup
The MCP server is implemented in Python and requires Python 3.13+
Fatebook MCP Server
A Model Context Protocol (MCP) server that provides integration with Fatebook, a prediction tracking platform. This server allows AI assistants like Claude to create, manage, and track predictions directly through MCP.
Features
- Create Questions: Make predictions with forecasts (0-100% probability)
- List Questions: View your predictions with filtering options
- Get Question Details: Retrieve comprehensive information about specific questions
- Add Forecasts: Update your predicted probabilities on existing questions
- Add Comments: Add commentary to track your reasoning
- Resolve Questions: Mark questions as resolved (YES/NO/AMBIGUOUS)
- Edit Questions: Update question titles, resolve dates, and notes
- Delete Questions: Remove questions you no longer need
- Count Forecasts: Track your forecasting activity
Testing and Compatibility
This MCP server has been tested on Mac OS X 14.5 with:
- Claude Code (CLI)
- Claude Desktop
- MCP Inspector
As far as I know, it doesn't violate any MCP protocol requirements but given that MCP is an evolving protocol, certain features or MCP clients may not be fully supported. If you encounter issues with other MCP implementations, please report them as GitHub issues.
Installation
Prerequisites
- Python 3.13 or higher
- uv (recommended) or pip
- A Fatebook account and API key
Getting your Fatebook API Key
Option 1: Install from PyPI (Recommended)
Install the package directly from PyPI:
Or using uv:
Option 2: Install from Source
For development or the latest features:
- Clone the repository:
- Install dependencies using uv:
Quick Start
Running Directly with uvx
The easiest way to test the server after installation is with uvx:
The server will start and wait for MCP client connections via stdio. Use Ctrl+C to stop it.
Usage with Claude Desktop and Claude Code
Claude Desktop
Add one of the following configurations to your Claude Desktop configuration file:
macOS
Location: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows
Location: %APPDATA%\Claude\claude_desktop_config.json
Option 1: Using PyPI Package (Recommended)
Note: If Claude Desktop can't find uvx
, use the full path instead:
Find your uvx path with: which uvx
Option 2: Development/Source Installation
Replace /path/to/fatebook-mcp
with the actual path to where you cloned this repository.
Claude Code (CLI)
For Claude Code, you can add this server in several ways:
Option 1: Using PyPI Package (Recommended)
Option 2: Import from Claude Desktop
If you already have this configured in Claude Desktop, you can import those settings:
Option 3: Project-specific configuration
Create a .mcp.json
file in your project:
Using PyPI package:
If uvx
is not found, use the full path (find with which uvx
):
Using development/source installation:
Replace /path/to/fatebook-mcp
with the actual path to where you cloned this repository.
Development & Testing
Additional setup
Beyond the above setup that's needed even just to use the MCP server, running the MCP server as standalone and running its tests require fetching your API key and exporting it in your environment.
You can do this either by exporting it directly:
Or create a .env
file:
Running integration tests
Run the integration tests to verify your setup:
This will test all available endpoints and confirm the server is working correctly.
Note: These tests will only succeed if you have the right API key for the test user. If you're developing such that you need to run these tests, for now email me (the author).
Running the Server Locally
Using uvx (recommended):
Using uv run with module:
The server will start and wait for MCP client connections.
Testing and Compatibility
This MCP server has been tested with:
- Claude Code (CLI)
- Claude Desktop
- MCP Inspector
As MCP is an evolving protocol, certain features or MCP clients may not be fully supported. If you encounter issues with other MCP implementations, please report them as GitHub issues.
Troubleshooting
uvx Command Not Found
If Claude Desktop shows "command not found" errors for uvx
:
- Find your uvx path:
- Use the full path in your configuration:
- Common uvx locations:
- macOS:
/Users/yourusername/.local/bin/uvx
- Linux:
/home/yourusername/.local/bin/uvx
- macOS:
API Key Issues
- Ensure your API key is valid and active
- Check that the API key is properly set in environment or config
- Try providing the API key directly in the request
Connection Issues
- Verify the server path in your Claude Desktop config is correct
- Check that Python 3.13+ is installed and accessible
- Ensure uv dependencies are properly installed with
uv sync
Question Format
- Dates should be in YYYY-MM-DD format
- Forecast values must be between 0.0 and 1.0
- Question IDs are provided when creating questions
Examples
Claude Desktop
Creating and tracking a prediction:
Reviewing your predictions:
Claude Code
API Documentation
For more details about the Fatebook API, see:
Contributing
Contributions are welcome! Please feel free to submit issues or pull requests.
Support
For issues with:
- This MCP server: Open an issue on GitHub
- Fatebook API: Contact Fatebook support
- MCP/Claude Desktop: See MCP documentation
remote-capable server
The server can be hosted and run remotely because it primarily relies on remote services or has no dependency on the local environment.
Tools
Enables AI assistants to create, manage, and track predictions on Fatebook, a prediction tracking platform. Supports making forecasts, updating probabilities, resolving questions, and managing prediction history through natural language.
Related MCP Servers
- -securityFlicense-qualityEnables AI assistants to interact with Metabase, providing access to dashboards, questions, databases, and tools for executing queries and viewing data through natural language.Last updated -
- AsecurityAlicenseAqualityEnables AI assistants to access and analyze financial data including stock information, company fundamentals, and market insights through the Financial Modeling Prep API.Last updated -25336TypeScriptApache 2.0
- AsecurityAlicenseAqualityA bridge between AI assistants and the Paradex perpetual futures trading platform that enables retrieving market data, managing trading accounts, placing orders, and monitoring positions.Last updated -261MIT License
- AsecurityAlicenseAqualityServer that enables AI assistants to interact with Storybook design systems. Extract component HTML, analyze styles, and help with design system adoption and refactoring.Last updated -86319MIT License