sequential-thinking-mcp
Remote MCP server for dynamic and reflective problem-solving through structured thinking. Enables step-by-step reasoning with branching, revision tracking, and session management.
README
š§ Sequential Thinking MCP Server
A Remote MCP Server implementation for dynamic and reflective problem-solving through structured thinking. Built with the official @vercel/mcp-adapter for seamless Vercel deployment with Streamable HTTP transport.
⨠Features
- š Streamable HTTP Transport - Modern, efficient transport (recommended by MCP spec March 2025)
- š SSE Support - Backward compatibility with Server-Sent Events
- š¦ Vercel Native - One-click deployment with
@vercel/mcp-adapter - š Branching Logic - Explore alternative thinking paths
- š Revision Tracking - Refine and improve previous thoughts
- š¾ Session Management - Persistent thinking across interactions
- š·ļø Auto-Classification - Automatic thought type detection
š Quick Deploy
Deploy to Vercel (Recommended)
After deployment, your MCP server will be available at:
- Streamable HTTP:
https://your-app.vercel.app/mcp - SSE:
https://your-app.vercel.app/sse
Local Development
# Clone the repository
git clone https://github.com/sheikhcoders/sequential-thinking-mcp.git
cd sequential-thinking-mcp
# Install dependencies
npm install
# Build
npm run build
# Run stdio mode (for MCP clients)
npm start
# Run HTTP mode (for development)
npm run start:http
š” Transport Modes
Streamable HTTP (Recommended)
The latest MCP transport specification. Eliminates persistent connections for better scalability.
POST https://your-app.vercel.app/mcp
Server-Sent Events (SSE)
Legacy transport for backward compatibility.
GET https://your-app.vercel.app/sse
Standard IO (stdio)
For local MCP clients like Claude Desktop.
node dist/index.js
š§ Configuration
Claude Desktop
Add to ~/.claude/claude_desktop_config.json:
{
"mcpServers": {
"sequential-thinking": {
"url": "https://your-app.vercel.app/mcp"
}
}
}
Cursor
Add to .cursor/mcp.json:
{
"mcpServers": {
"sequential-thinking": {
"url": "https://your-app.vercel.app/mcp"
}
}
}
Cline (Streamable HTTP)
Add to cline_mcp_settings.json:
{
"mcpServers": {
"sequential-thinking": {
"command": "npx",
"args": ["mcp-remote", "https://your-app.vercel.app/mcp"],
"transportType": "Streamable HTTP"
}
}
}
For stdio (Local)
{
"mcpServers": {
"sequential-thinking": {
"command": "npx",
"args": ["sequential-thinking-mcp"]
}
}
}
š ļø Available Tools
sequential_thinking
The main tool for step-by-step problem solving with dynamic thought management.
Parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
thought |
string | ā | Your current thinking step |
nextThoughtNeeded |
boolean | ā | Whether another thought is needed |
thoughtNumber |
number | ā | Current thought number (1-indexed) |
totalThoughts |
number | ā | Estimated total thoughts needed |
isRevision |
boolean | ā | Whether this revises previous thinking |
revisesThought |
number | ā | Which thought number is being revised |
branchFromThought |
number | ā | Create branch from this thought |
branchId |
string | ā | Branch identifier |
sessionId |
string | ā | Session ID for persistence |
get_thinking_summary
Get a comprehensive summary of a thinking session.
list_thinking_sessions
List all available thinking sessions.
switch_thinking_branch
Switch between different thinking branches.
complete_thinking_session
Mark a session as completed with optional final conclusion.
š Thought Types
Thoughts are automatically classified:
| Type | Detected When |
|---|---|
question |
Contains ?, starts with what/how/why |
observation |
Contains "I notice", "I see", "observe" |
hypothesis |
Contains "perhaps", "maybe", "hypothesis" |
verification |
Contains "verify", "test", "check" |
insight |
Contains "insight", "realize", "aha" |
conclusion |
Contains "therefore", "in conclusion" |
refinement |
Contains "refine", "improve", "better" |
reflection |
Contains "reflect", "thinking about" |
analysis |
Default for analytical statements |
šļø Project Structure
sequential-thinking-mcp/
āāā api/
ā āāā [transport]/
ā āāā route.ts # Vercel serverless handler
āāā src/
ā āāā index.ts # CLI entry point (stdio mode)
ā āāā server.ts # MCP stdio server
ā āāā http-transport.ts # Express HTTP server
ā āāā thinking-session.ts # Session management
ā āāā types.ts # TypeScript definitions
āāā package.json
āāā tsconfig.json
āāā vercel.json # Vercel configuration
āāā README.md
š Production Considerations
Session Persistence
For production with persistent sessions, integrate Vercel KV or Upstash Redis:
import { kv } from '@vercel/kv';
// Store session
await kv.set(`session:${sessionId}`, session);
// Retrieve session
const session = await kv.get(`session:${sessionId}`);
Authentication
Add OAuth or API key authentication for production:
// In your handler
const authHeader = request.headers.get('Authorization');
if (!validateToken(authHeader)) {
return new Response('Unauthorized', { status: 401 });
}
š API Reference
Health Check
curl https://your-app.vercel.app/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc": "2.0", "method": "initialize", "id": 1}'
Call Tool
curl https://your-app.vercel.app/mcp \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "sequential_thinking",
"arguments": {
"thought": "Let me analyze this problem step by step",
"nextThoughtNeeded": true,
"thoughtNumber": 1,
"totalThoughts": 5
}
},
"id": 1
}'
š Resources
š License
MIT License - see LICENSE for details.
š¤ Contributing
Contributions welcome! Please read our contributing guidelines and submit PRs.
Built with ā¤ļø using the Model Context Protocol
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.