Sequential Thinking MCP server
Last verified: 2026-09
A scratchpad that forces the model to break work into steps.
Official reference Developer tools No API key Claude Desktop, Cursor, Windsurf, Claude Code, VS Code
What it does
A scratchpad tool that makes the model write a chain of thoughts before answering. No external data, no secrets. Useful when you want to see the plan; easy to overuse.
It sits in Developer tools: Split this category in two.
Good for
- Force a multi-step plan the host will show as a thought list.
- Break a refactor into ordered steps you can reject mid-way.
- Debug a client by watching it record thoughts — then remove the server.
- Skip it for lookup work; hosts already reason.
Tools
- sequentialthinking — Append or revise a thought step.
Config (Claude Desktop / Cursor / Windsurf)
Paste the JSON below. Same mcpServers shape. Replace placeholder paths and secrets.
Windows paths look like C:\\Users\\you\\project, not /path/to.
Host-side steps →
Claude Desktop
| OS | Config file |
|---|---|
| macOS | ~/Library/Application Support/Claude/claude_desktop_config.json |
| Windows | %APPDATA%\Claude\claude_desktop_config.json
(usually C:\Users\<you>\AppData\Roaming\Claude\) |
| Linux | ~/.config/Claude/claude_desktop_config.json |
Cursor
| Scope | macOS / Linux | Windows |
|---|---|---|
| This project | .cursor/mcp.json in the repo root | |
| This user | ~/.cursor/mcp.json |
%USERPROFILE%\.cursor\mcp.json |
Windsurf
| OS | Config file |
|---|---|
| macOS / Linux | ~/.codeium/windsurf/mcp_config.json |
| Windows | %USERPROFILE%\.codeium\windsurf\mcp_config.json |
{
"mcpServers": {
"sequential_thinking": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-sequential-thinking"
]
}
}
}
Windsurf remote MCP: use serverUrl instead of url if the block below is HTTP.
One-liner: npx -y @modelcontextprotocol/server-sequential-thinking
Secrets it wants: none
How to get started
- Add the JSON. No env.
- Ask a multi-step question and tell the model to use sequential thinking.
- Turn it off if the extra steps just slow simple tasks.
Access risk
None beyond wasting tokens. Thoughts can echo secrets you already pasted in the chat.
When to skip it
Skip it for lookup tasks (Fetch, GitHub). Hosts already do chain-of-thought without this server.
Instead: Everything.
Vs alternatives
| Server | Official? | Needs a secret? | Best for |
|---|---|---|---|
| Sequential Thinking | Yes | No | A scratchpad that forces the model to break work into steps. |
| Everything | Yes | No | Reference server that exercises the whole MCP protocol. |
More in Developer tools
FAQ
Does this make the model smarter?
Sometimes it makes the plan readable. It does not add tools to the world. Skip it for 'what is in this file'.
Do I leave it installed?
No need. It is a scratchpad. Install for a hard planning session if you like the UI, then drop it.
Official?
Yes — reference server. No secrets.
Sequential Thinking or Everything?
Everything is a protocol test bed. Sequential Thinking is a reasoning scratchpad. Neither talks to your systems.