deliberate-reasoning-engine
MCP server that transforms linear AI reasoning into structured, auditable thought graphs
claude mcp add --transport stdio haasonsaas-deliberate-reasoning-engine npx -y deliberate-reasoning-engine
How to use
To use DRE, connect to the MCP server via your MCP client and call the provided tools. The available tools include log_thought for adding a structured thought with a semantic type and optional dependencies, get_thought_graph to fetch the current DAG of thoughts (with options for full or summary formats), and invalidate_assumption to mark an assumption as invalid and cascade updates to dependent thoughts. A typical flow starts with logging an objective, then adding hypotheses and assumptions with explicit dependencies, followed by sub_problems and actions as the reasoning progresses. You can also retrieve the graph to visualize relationships and ensure there are no cycles or orphaned thoughts. The example workflow demonstrates creating an objective, adding a hypothesis with a dependency on the objective, introducing an assumption with a link to the hypothesis, decomposing into a sub_problem, and invalidating an assumption to propagate changes.
How to install
Prerequisites:
- Node.js v18+ and npm installed
- Internet access to install packages
Installation steps:
- Install the DRE package globally (recommended):
npm install -g deliberate-reasoning-engine
- (Optional for development) Clone the repository and install dependencies:
git clone https://github.com/haasonsaas/deliberate-reasoning-engine.git
cd deliberate-reasoning-engine
npm install
- Build (if you cloned the repo and plan to run from source):
npm run build
- Run the server (example for development):
npm run dev
Notes:
- If you installed globally, you can run the server directly via the command provided by the package.
- Ensure Node.js version compatibility (Node.js 18+ is recommended as per README).
Additional notes
Tips and common issues:
- When configuring in Claude Desktop or other MCP clients, use an absolute path to the server binary if you run locally from dist/index.js.
- If you encounter module-not-found errors, run npm install and ensure you’re using Node.js 18+.
- The MCP config can be adjusted to run via npx with the package name (deliberate-reasoning-engine) or via a local dist/index.js path if you build from source.
- Keep an eye on roadmap features such as session persistence and graph visualization exports for enhanced workflows.
- For troubleshooting, verify that the server is reachable by the MCP client and that the command/arguments match the deployment method you chose (npx vs node path vs development server).
- The npm package name is deliberate-reasoning-engine; you can replace with a local path or a specific executable if needed.
Related MCP Servers
mcp-telegram
MCP Server for Telegram
Remote
A type-safe solution to remote MCP communication, enabling effortless integration for centralized management of Model Context.
recall
Persistent cross-session memory for Claude & AI agents. Self-host on Redis/Valkey, or use the managed SaaS at recallmcp.com.
boilerplate
TypeScript Model Context Protocol (MCP) server boilerplate providing IP lookup tools/resources. Includes CLI support and extensible structure for connecting AI systems (LLMs) to external data sources like ip-api.com. Ideal template for creating new MCP integrations via Node.js.
mobile
A Model Context Protocol (MCP) server that provides mobile automation capabilities.
docmole
Dig through any documentation with AI - MCP server for Claude, Cursor, and other AI assistants