sktime-mcp

The Semantic Engine for Time-Series Forecasting with Large Language Models.

sktime-mcp is a Model Context Protocol (MCP) server that exposes the full power of the sktime ecosystem to AI assistants. It provides a stateful runtime for discovering, instantiating, and executing complex time-series forecasting workflows.

Getting Started

• Installation
  • Prerequisites
  • Standard Installation
  • Developer Installation
  • Verification
• Quickstart
  • 1. Start the Server
  • 2. Connect to a Client
  • 3. Verify Functionality
  • 4. Run Your First Forecast
  • Next Steps
• Core Concepts
  • Collaborative Model Discovery
  • Stateful Interaction (Handles)
  • Asynchronous Background Jobs
  • Safety and Reproducibility

User Guide

• Data Management
  • Loading Your Data
  • Working with Handles
  • Formatting and Cleaning
  • Managing Resources
  • Built-in Demo Datasets
• Forecasting Workflows
  • Model Discovery
  • The Fit-Predict Workflow
  • Evaluation
  • Best Practices for Users
• Pipelines and Composition
  • Describing a Pipeline
  • Validation
  • Creating and Using Pipelines
  • Common Use Cases
• Asynchronous Operations
  • When to Use Background Tasks
  • How to Start a Background Task
  • Monitoring Progress
  • Managing Jobs
• Persistence and Export
  • Saving and Loading Your Models
  • Exporting Your Workflow as Code
  • Important Considerations

Developer Guide

• Architecture
  • High-Level Overview
  • Core Components
  • Data Flow: The “Handle” Pattern
  • Concurrency Model
  • Security Considerations
• Contributing Guide
  • Development Setup
  • Code Quality Standards
  • Testing
  • Project Structure
  • Pull Request Process
• API Reference
  • EstimatorNode
  • RegistryInterface
  • TagResolver
  • CompositionValidator
  • Executor
  • HandleManager
  • Server
  • Tools
  • Registry
  • Runtime
  • Composition

Indices and tables

• Index

• Module Index

• Search Page