kansei-link/kansei-mcp-server

Overview

A local-first Model Context Protocol (MCP) server that gives AI agents direct access to over 11,000 SaaS services and 200 workflow recipes. It runs on the developer's machine, works with Claude Code, Cursor, Cline, Zed, and Windsurf, and claims 89-97% token savings by reducing redundant data passed to language models.

Best for

Best for
Developers using MCP-compatible coding tools who want local access to a wide range of SaaS services with potential token cost savings.

Use cases

• Connecting AI coding agents to external APIs like Slack, Notion, or GitHub via a local MCP server
• Running predefined workflow recipes to automate multi-step tasks across different SaaS tools
• Reducing token consumption in agentic workflows when using Claude Code or Cursor

Notes

A local-first Model Context Protocol (MCP) server that gives AI agents direct access to over 11,000 SaaS services and 200 workflow recipes. It runs on the developer’s machine, works with Claude Code, Cursor, Cline, Zed, and Windsurf, and claims 89-97% token savings by reducing redundant data passed to language models.

1 stars on GitHub. Last updated 2026-07-13. Licensed MIT.

Use cases

• Connecting AI coding agents to external APIs like Slack, Notion, or GitHub via a local MCP server
• Running predefined workflow recipes to automate multi-step tasks across different SaaS tools
• Reducing token consumption in agentic workflows when using Claude Code or Cursor

Pros

• Large built-in catalog of 11,000+ SaaS services reduces setup time
• Local-first architecture preserves privacy and works offline
• Compatible with multiple popular AI coding tools out of the box

Cons

• Very early-stage project with only 1 GitHub star and minimal community adoption
• Token savings claims of 89-97% are unverified and likely vary by workflow complexity
• Requires local configuration and understanding of MCP to integrate effectively

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.