How Brainstorm Reactor thinks.
The quality of AI output depends on process design, not prompt engineering skill. Brainstorm Reactor shifts control: away from prompting, toward designing cognitive recipes.
One stream.
Five composable layers.
Every API call builds a system prompt from five independent layers. The same recipe step produces fundamentally different results depending on persona, context, and method. Library of 200+ frameworks.
Persona
Cognitive stance — analytical, creative, adversarial, or curatorial
Method
One of 200+ thinking methods, injected as a structured instruction
Context
Pinned content + pruned history (full, last_step, or none)
Slot
Model selection — LOGIC, CREATIVE, or SEARCH per step
Output
Structured nodes, streamed in real-time with transparent reasoning
Three slots.
Heterogeneous composition.
Each recipe step declares the required cognitive slot. Analytical work goes to reasoning models, research to web models, synthesis to generative models — all in one workflow.
Analytical Reasoning
Reasoning models (o1, o3-mini, Gemini Flash Thinking)
TRIZ analysis, contradiction detection, scoring, validation, adversarial audit
Ideation & Synthesis
Generative models (Claude Sonnet, GPT-4o, Llama 4)
Brainstorming, storytelling, concept synthesis, solution architecture, copywriting
Evidence-Based Research
Web models (Perplexity Sonar, Gemini with grounding)
Patent research, competitive analysis, market research, state of the art
Six ways to orchestrate thinking.
From selective amnesia to parallel swarms — every recipe is built on composable execution patterns.View all recipes
Shadow Pass & Selective Amnesia
Sequential steps with precise context scopes. Using 'last_step' context mode allows the engine to completely forget original user prompt bias (selective amnesia) for a clean synthesis.
e.g. TRIZ Express, Elevator Pitch
Parallel Swarm & Synthesis
Simultaneous adversarial execution of multiple agents with divergent directives. A centralized synthesizer compiles their outputs from a compressed XML matrix.
e.g. MAD Engine, Ergodic Hive
Autonomous Critique & Revise (Aletheia)
An autonomous feedback loop where an auditor step evaluates outputs against constraints, running backward jumps with pruned context (Context Collapse) until passing.
e.g. Aletheia Engine, TRIZ v9 MAX
User Decision Branching (Router)
Workflow pauses at defined decision points. Using keyword matching on quick-reply buttons (with optional D2 Split Router), the user directs the flow dynamically along specialized branches.
e.g. Dilemma Decoder, Branching Demo
Unconditional Forward Jump
A converging execution flow where multiple parallel branches or slots automatically jump forward to a single consolidation step using empty routes.
e.g. Split Router, Configurator Suite
Epistemic Anchor (Auto-Pin)
Pins foundational outputs to the persistent '<pinned_context>' layer. Subsequent steps can access this baseline directly, even when operating with selective amnesia.
e.g. Data Baseline, Core Compiler
Composable Patterns
Patterns can be freely combined. Real recipes are hybrids:
- C+BAutonomous swarm — parallel agents + self-correction
- D+BBranching swarm — user decision → specialized swarm
- C+DCopilot / Autopilot — user switches between manual routing and autonomous loop
Glass Box UI.
Every AI response is fully transparent. No black box. See what the model thought before it wrote — inspect structured data and debug raw output — all in real-time.
Agent labels show which specialist is active during multi-step recipes. The Cognitive Protocol reveals the model's internal reasoning process before a single word of output appears.
Formatted output — clean, structured, actionable
Native reasoning trace from the model's thought process
Structured node data — parseable, exportable
Unprocessed model output for debugging
TRIZ v8 Orchestrator.
5 specialized agents. 3 model slots. 2 context modes. The flagship recipe demonstrates full orchestrator capacity.
Decompose the problem into components, functions, and contradictions.
Apply the TRIZ contradiction matrix. Generate solution directions.
Research existing solutions, patents, and analogous domains.
Synthesize all insights into concrete, actionable concepts.
Identify weaknesses, risks, and unintended consequences.
Context Pruning
Steps 2 & 3 use last_step — the model focuses on the distilled output, not the full history.
Slot Diversity
Analysis → reasoning models. Research → web models. Synthesis → creative models. No single model does everything.
Adversarial Close
Step 5 sees the full context but uses a reasoning model — maximum scrutiny on the entire proposal.
Autonomous Diagnostics & Policy Cockpits in the Field.
Socratic requirements intake and hierarchical directive governance operate on the exact same deterministic multi-model orchestration as complex enterprise SaaS audits — empirically validated via Swim Coach Pro.
Why does a cognitive process engine built for enterprise architecture prove its reliability at the edge of the pool? Because high-stakes organizational systems obey universal principles: Unstructured user needs require dynamic Socratic elicitation, while institutional supervisors must deterministically bound autonomous agent execution. Our white-labelable platform demonstrates this dual-tier architecture in the independent /swimcoach reference dropzone, packaging the same server-side engine for vertical deployments across industry, healthcare, and strategic consulting.
Dynamic Requirements Intake
Active conversational dialectic extracting unstructured constraints, fatigue indicators, and operational boundaries without rigid forms.
Supervisor Policy Envelopes
Two-tier directive control: Institutional policies strictly bound the executing agent to defined workload ceilings and safety corridors.
Validated Action Blueprints
Deterministic compilation of diagnostic parameters into verified 6-block execution plans with wearable sync and multimodal telemetry.
Decoupling Reasoning from Form.
Forcing AI models to deduce logic and apply prose formatting in the same forward pass induces a severe style tax (-12.4% accuracy). The Brainstorm Reactor isolates the logical search space entirely from surface formatting.
Heterogeneous Model Economy
Fast SLMs handle parsing, while specialized reasoning engines resolve contradictions, reducing operational inference costs by 70–85% compared to monolithic models.
Model-Agnostic Hardening
PH-SRTP frees native reasoning models from style constraints in their scratchpads while equipping standard models with multi-pass dialectical rigor.
Interactive Observability
Zero black-box blind spots. Every reasoning vessel streams intermediate state vectors in real time, giving human decision-makers live steering control.
Impact of Style Tax on Strategic Depth (PlanJudge 72B)
Evaluated on GPT-5.6 Sol against the CCA Strategic Transformation Paradox across 5 anchored dimensions (Structural Depth, Analytical Rigor, Divergent Breadth, Actionability, Anti-Sycophancy).
Why cognitive architecture beats mega-prompts.
In March 2026, the Kimi team published the paper Attention Residuals — a mathematical proof of a problem that Brainstorm Reactor already solves architecturally at the process control level.
PreNorm Dilution
In deep networks, residual connections accumulate. Essential early information gets diluted by noise from middle layers — deeper layers lose access to the original signals.
Context Contamination
In standard chats, context accumulates message by message. By the time a model reaches step 6, the original problem is diluted in the noise of iterative intermediate steps.
Skip Connections → Epistemic Anchor
Kimi: Later layers skip accumulated noise and directly access early, clean layers (Attention Residuals).
Brainstorm Reactor: The Epistemic Anchor preserves extracted facts from step 1 via auto-pinning. Later agents with context_mode: last_step work only with the distilled output — the ground truth remains directly accessible via <pinned_context>.
Block Compression → Swarm Synthesis
Kimi: Layers are grouped into blocks and compressed into a single vector. Later layers see only the clean summary, not the raw individual steps.
Brainstorm Reactor: When parallel agents work (Pattern F), a synthesizer step compresses the outputs into a dense XML aggregate. Subsequent steps process only this node — not the individual agent responses.
Deep & Narrow → Microsteps
Kimi: Networks with Attention Residuals reach their optimum with deeper, narrower architectures — shallow networks stagnate.
Brainstorm Reactor: Many tightly focused steps (deep & narrow) beat few overloaded mega-prompts (shallow & wide). When the engine handles navigation and context pruning, the model can concentrate 100% of its parameters on pure transformation.
Intelligence in complex systems doesn't emerge from endlessly accumulating data, but from targeted noise reduction — regardless of whether the system is a neural network or a cognitive pipeline.
Source: Kimi / Moonshot AI, Attention Residuals (March 2026) · Inspect benchmark results →
Unabhängig. Souverän. DSGVO-nativ.
Die Brücke zwischen maximaler Modellfreiheit und kompromissloser Datensicherheit. Der Brainstorm Reactor kapselt seine kognitive Leistung serverseitig und schützt geschäftskritische Daten vor unbefugtem Abfluss.
EU-native Infrastruktur
Die gesamte Kernplattform und alle Datenbanken werden ausnahmslos in der Europäischen Union (Deutschland und Irland) betrieben.
Zero Data Retention
Verarbeitung ausschließlich über geschäftliche B2B-Schnittstellen. Eingaben und hochgeladene Dokumente sind vertraglich streng vor dem KI-Modelltraining geschützt.
Datenschutz-Gateway (opt-in)
Optionale Echtzeit-Maskierung. Filtert personenbezogene Daten (PII) automatisch auf EU-Servern heraus, bevor Anfragen verarbeitet werden.
Ausfallsicheres Routing
Provider-unabhängiges Modell-Routing (OpenAI, Anthropic, Google, DeepSeek). Fällt ein KI-Anbieter aus, läuft deine Prozess-Pipeline nahtlos weiter.
Methods think ahead. Think along. Think further.
200+ methods. 138+ recipes. Three model slots — ready to steer.
Open thinking space