Reasoning

Reasoning is a specialized AI tool platform that addresses a critical failure mode in modern AI assistants: their tendency to validate user premises without critical challenge. It offers a suite of four cognitive tools—Structured Reflection, Sequential Thinking, Context Switcher, and Decision Matrix—that bind an AI agent to a rigorous, stage-gated process. This forces the AI to articulate problems, consider multiple perspectives, and evaluate options with weighted criteria before providing answers. Each tool runs in an isolated session to prevent context pollution in your primary workflow.

How to use Reasoning

Reasoning operates via the Model Context Protocol (MCP). To use it, subscribe to get API credentials, then add the tool endpoints to your MCP-compatible client configuration (such as Claude Desktop or Cursor). Once configured, you can invoke the tools naturally within your existing AI assistant, for example by prompting: "Start a Decision Matrix session for this database choice." The tools will then guide the AI through a structured reasoning process.

Core Features of Reasoning

• Structured Reflection: An interactive session that asks clarifying questions to force you to articulate the core problem, breaking through mental blocks.
• Sequential Thinking: Implements stage gates (e.g., define problem before researching, research before analyzing) to prevent jumping to solutions and catch flawed reasoning.
• Context Switcher: Simultaneously evaluates a problem from multiple stakeholder perspectives (technical, business, user, risk, ops) using templates for architecture, security, and debugging.
• Decision Matrix: Structures complex decisions with weighted criteria, scored options, and sensitivity analysis to move beyond gut feelings validated by a helpful AI.
• Isolated Sessions: All reasoning occurs in clean, separate contexts, ensuring insights are extracted without polluting your main AI conversation thread.

Use Cases for Reasoning

This tool is primarily for developers and technical founders using AI coding agents (like Claude Code, Cursor, Windsurf). It's essential when an agent is stuck in a loop (iteration 30 of the wrong approach) to break the cycle with a fresh context. It's also critical for making high-stakes, difficult-to-reverse decisions (architecture, vendor, hire) where you need more than superficial AI validation—specifically, documented justification and scrutiny-proof reasoning.

Advantages of Reasoning

Reasoning's key advantage is that it provides a "contract" rather than a "suggestion" to the AI. Unlike prompt engineering where a model can ignore a "be critical" instruction, these tools use actual tool calls (e.g., `set_weights`, `score_option`) that the agent must invoke, making the reasoning auditable and observable. It's a solo-founder-built, practical tool focused on clear thinking, offering a flat-rate subscription with no usage metering to encourage use when needed.

Pricing and Plans

Reasoning offers a simple, flat-rate pricing model with a 14-day free trial. The Monthly plan costs $20/month and provides full access to all reasoning tools, generous usage limits, session history, exports, and email support. The Annual plan is $180/year (a 25% saving), offering the same features plus priority support. Both plans are cancel anytime from account settings with no usage overage charges.