ThoughtMCP

# Parallel Reasoning Streams Guide This guide explains how ThoughtMCP's parallel reasoning streams work together to provide comprehensive analysis from multiple perspectives simultaneously. ## Overview The `think_parallel` tool processes problems through four specialized reasoning streams that work simultaneously, each bringing a unique perspective to the analysis. This mimics how human experts often approach complex problems by considering multiple angles at once. ## The Four Reasoning Streams ### 🔬 Analytical Stream **Role**: Logical, data-driven evaluation and systematic analysis **What it does:** - Breaks down problems into logical components - Evaluates evidence and data systematically - Builds structured argument chains - Performs quantitative analysis when possible - Identifies cause-and-effect relationships **Best for:** - Technical problems requiring logical reasoning - Decisions with measurable outcomes - Situations with clear data and metrics - Problems needing systematic evaluation **Example Output:** ``` "Based on the financial data, Option A shows 23% higher ROI over 3 years. The technical architecture supports 10x current load with 99.9% uptime. Risk analysis indicates 15% probability of major issues vs 35% for alternatives." ``` **When Analytical Stream Leads:** - Engineering decisions - Financial analysis - Performance optimization - Data-driven choices ### 🎨 Creative Stream **Role**: Innovative approaches and unconventional alternatives **What it does:** - Generates novel solutions and approaches - Challenges conventional thinking - Explores "what if" scenarios - Combines ideas from different domains - Identifies breakthrough opportunities **Best for:** - Innovation challenges - Stuck or recurring problems - Need for differentiation - Breakthrough solutions required **Example Output:** ``` "What if we flip the problem? Instead of reducing costs, increase value perception. Consider a freemium model like Spotify - free tier drives premium conversions. Could we partner with competitors to create industry standards that benefit everyone?" ``` **When Creative Stream Leads:** - Product innovation - Marketing strategies - Business model design - Competitive differentiation ### ⚠️ Critical Stream **Role**: Risk assessment, assumption challenging, and quality control **What it does:** - Identifies potential risks and downsides - Challenges assumptions and biases - Evaluates argument quality and evidence - Considers what could go wrong - Provides skeptical perspective **Best for:** - High-stakes decisions - Risk assessment scenarios - Quality assurance needs - Bias detection requirements **Example Output:** ``` "Key assumption: 'Users will adopt new interface easily' - but our last UI change had 40% user complaints. Risk: Implementation timeline assumes no major bugs, but similar projects averaged 2.3x time overruns. Consider: What if our main competitor launches first?" ``` **When Critical Stream Leads:** - Investment decisions - Safety-critical systems - Regulatory compliance - Crisis management ### 🔄 Synthetic Stream **Role**: Integration, holistic perspective, and consensus building **What it does:** - Integrates insights from other streams - Finds common ground and reconciles differences - Provides holistic, big-picture perspective - Balances competing priorities - Synthesizes final recommendations **Best for:** - Complex multi-faceted problems - Stakeholder alignment needs - Strategic decision making - Integration challenges **Example Output:** ``` "Integrating all perspectives: The analytical data supports Option A, creative insights suggest hybrid approach, critical analysis highlights timeline risks. Recommendation: Phased implementation starting with Option A core features, incorporating creative enhancements in phase 2, with critical risk mitigations built into each phase." ``` **When Synthetic Stream Leads:** - Strategic planning - Complex negotiations - Multi-stakeholder decisions - System integration ## How Streams Work Together ### Real-Time Coordination Streams don't work in isolation - they actively coordinate and share insights: **Information Sharing:** - Analytical stream shares data findings with Creative stream - Creative stream's ideas are evaluated by Critical stream - Critical stream's concerns inform Analytical risk models - Synthetic stream integrates all perspectives continuously **Conflict Resolution:** - When streams disagree, they engage in structured dialogue - Evidence is weighed and assumptions are challenged - Compromises and hybrid solutions are explored - Final synthesis balances all valid perspectives **Consensus Building:** - Streams identify areas of agreement - Disagreements are clearly articulated with reasoning - Alternative approaches are developed when consensus isn't possible - Confidence levels reflect the degree of stream alignment ### Example: Coordination in Action **Problem**: "Should we migrate our monolithic application to microservices?" **Analytical Stream**: "Migration will cost $500K, take 8 months, but reduce deployment time from 2 hours to 15 minutes and improve scalability by 300%." **Creative Stream**: "What if we use a strangler fig pattern? Gradually extract services while keeping the monolith running. Or consider serverless functions for new features only." **Critical Stream**: "Risk: Team has no microservices experience. Similar migrations have 60% failure rate. What if we can't maintain data consistency across services?" **Coordination Point**: Creative stream's strangler fig idea addresses Critical stream's risk concerns while maintaining Analytical stream's benefits. **Synthetic Stream**: "Hybrid approach: Start with strangler fig pattern for low-risk modules, build team expertise gradually, maintain monolith for core functions until confidence is high." ## Stream Selection and Emphasis ### Automatic Stream Weighting ThoughtMCP automatically adjusts stream emphasis based on problem characteristics: **Technical Problems**: Analytical stream gets higher weight **Innovation Challenges**: Creative stream leads **High-Risk Decisions**: Critical stream emphasis increases **Complex Integration**: Synthetic stream coordination intensifies ### Manual Stream Configuration You can influence stream behavior through context: ```json { "tool": "think_parallel", "arguments": { "input": "Your problem...", "context": { "emphasis": "creative", // Boost creative stream "risk_tolerance": "low", // Increase critical stream activity "innovation_priority": "high" // Creative + Analytical focus } } } ``` ## Understanding Stream Outputs ### Reading Stream Results Each stream provides: - **Core Analysis**: Main reasoning and conclusions - **Supporting Evidence**: Data, examples, and rationale - **Confidence Level**: How certain the stream is about its conclusions - **Key Insights**: Most important discoveries or realizations - **Recommendations**: Specific actions or approaches suggested ### Interpreting Coordination Look for these coordination patterns: **High Agreement**: All streams reach similar conclusions - High confidence in the recommendation - Clear path forward - Low risk of overlooking important factors **Productive Disagreement**: Streams disagree but identify why - Multiple valid approaches exist - Trade-offs are clearly articulated - Hybrid solutions often emerge **Unresolved Conflict**: Streams can't reach consensus - More information needed - Fundamental uncertainty exists - Consider gathering additional data ## Best Practices for Parallel Reasoning ### When to Use Parallel Reasoning **Ideal Scenarios:** - Strategic business decisions - Complex technical architecture choices - Investment and resource allocation - Crisis response planning - Innovation and product development **Less Suitable:** - Simple, well-understood problems - Time-critical decisions (use `think` in intuitive mode) - Personal, emotional decisions (use `think` with emotion enabled) - Routine operational choices ### Optimizing Input for Parallel Processing **Provide Rich Context:** ```json { "input": "Should we acquire StartupX for $50M?", "context": { "domain": "business-strategy", "urgency": 0.6, "complexity": 0.9, "stakeholders": ["board", "investors", "employees"], "constraints": ["budget_constraint", "timeline_constraint"], "strategic_goals": ["market_expansion", "technology_acquisition"] } } ``` **Frame Problems Clearly:** - State the decision or problem explicitly - Include relevant background information - Mention key constraints and requirements - Specify what success looks like ### Interpreting Results Effectively **Focus on Synthesis:** - The Synthetic stream's integration is usually most actionable - Look for how different perspectives complement each other - Pay attention to areas where streams agree strongly **Understand Disagreements:** - Disagreements often reveal important trade-offs - Critical stream concerns usually highlight real risks - Creative alternatives may solve apparent conflicts **Use Confidence Levels:** - High confidence across streams = strong recommendation - Low confidence = gather more information - Mixed confidence = consider phased or experimental approaches ## Advanced Coordination Features ### Stream Synchronization Streams synchronize at key points during processing: 1. **Initial Problem Analysis**: All streams analyze the problem independently 2. **Mid-Process Sync**: Streams share preliminary findings 3. **Conflict Detection**: Disagreements are identified and explored 4. **Final Integration**: Synthetic stream creates unified recommendation ### Dynamic Stream Adaptation Streams adapt their approach based on: - Problem complexity and domain - Available information quality - Time constraints and urgency - Risk tolerance and stakes - Previous stream interactions ### Quality Assurance Built-in quality checks ensure: - Each stream maintains its distinct perspective - Coordination doesn't lead to groupthink - Minority viewpoints are preserved - Evidence quality is maintained across streams ## Troubleshooting Parallel Reasoning ### Common Issues and Solutions **Streams Too Similar:** - Problem may be too simple for parallel processing - Try `think` or `analyze_systematically` instead - Add more context to differentiate perspectives **Overwhelming Output:** - Focus on the Synthetic stream's integration first - Look for key disagreements and their resolutions - Use the coordination insights to understand the process **Low Confidence Results:** - May indicate insufficient information - Consider gathering more data before deciding - Use `think_probabilistic` for uncertainty analysis **Conflicting Recommendations:** - Normal for complex problems with trade-offs - Look for hybrid approaches in Synthetic stream - Consider phased implementation strategies ## Examples by Domain ### Technology Decisions **Analytical**: Performance benchmarks, cost analysis, technical feasibility **Creative**: Novel architectures, innovative solutions, emerging technologies **Critical**: Security risks, scalability concerns, technical debt implications **Synthetic**: Balanced technical strategy with risk mitigation ### Business Strategy **Analytical**: Market data, financial projections, competitive analysis **Creative**: New business models, innovative partnerships, market disruption **Critical**: Market risks, competitive responses, execution challenges **Synthetic**: Integrated strategy balancing growth and risk ### Product Development **Analytical**: User data, feature usage, development costs **Creative**: Innovative features, new user experiences, market differentiation **Critical**: User adoption risks, technical complexity, resource constraints **Synthetic**: Product roadmap balancing innovation and feasibility ## Integration with Other Tools ### Combining with Systematic Analysis 1. `analyze_systematically` - Structure the problem 2. `think_parallel` - Explore multiple perspectives 3. `analyze_reasoning` - Quality check the conclusions ### Memory-Enhanced Parallel Reasoning 1. `recall` - Gather relevant past experience 2. `think_parallel` - Process with multiple perspectives 3. `remember` - Store insights from each stream ### Probabilistic Integration 1. `think_parallel` - Generate multiple scenarios 2. `think_probabilistic` - Assess probabilities and uncertainties 3. `decompose_problem` - Plan implementation based on insights --- _Want to see parallel reasoning in action? Check out our [Advanced Tools Examples](../examples/basic/advanced-tools.md#think_parallel---multi-stream-reasoning)._