mcp-think
A remote MCP server for structured sequential thinking, enabling users to create, revise, and manage sequences of steps with persistent state via Cloudflare Workers and Durable Objects.
README
mcp-think
A remote MCP server for structured sequential thinking, built for Cloudflare Workers.
Architecture
- Cloudflare Worker exposes
/mcpover Streamable HTTP. - MCP SDK v2 defines the tools.
- Durable Objects + SQLite persist each sequence.
- A sequence is addressed by an explicit
sequenceId; MCP protocol sessions do not carry application state.
This design follows the current Cloudflare guidance: new MCP servers use the v2/stateless handler, while application state for multi-step workflows lives behind an explicit Durable Object boundary. McpAgent is deprecated and is intentionally not used here.
Tools
start_sequence
Creates a sequence and returns a sequenceId.
add_step
Adds a structured step. Use this for summaries, assumptions, evidence, hypotheses, or decisions. The server is not intended to expose or store hidden chain-of-thought.
revise_step
Creates a new revision of a previous step while preserving the original record.
get_sequence
Returns the sequence metadata and all stored steps.
finish_sequence
Stores the final conclusion and marks the sequence as completed.
reset_sequence
Deletes the stored sequence data.
Local development
npm install
npm run typecheck
npm run dev
The local MCP endpoint is:
http://localhost:8787/mcp
Health check:
http://localhost:8787/health
Deploy
Authenticate Wrangler with your Cloudflare account and run:
npm run deploy
Cloudflare will provision the SQLite-backed Durable Object from the migration in wrangler.jsonc.
Example flow
start_sequence(goal)
↓
sequenceId
↓
add_step(sequenceId, ...)
↓
add_step(sequenceId, ...)
↓
revise_step(sequenceId, stepId, ...)
↓
get_sequence(sequenceId)
↓
finish_sequence(sequenceId, conclusion)
Security roadmap
The MVP intentionally has no authentication. Before exposing the endpoint publicly, add OAuth or another authentication layer and bind sequence ownership to the authenticated principal. Cloudflare's MCP guidance also recommends explicit Host/Origin validation and authentication rather than treating CORS as an auth mechanism.
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.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
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.