The Brainstorm Reactor operates as a cognitive exoskeleton for deep knowledge work. Rather than presenting an unstructured chat interface, the platform organizes its instruments across 4 distinct altitude levels, allowing users to navigate between focused one-step heuristics, multi-agent pipelines, tactile creative studios, and fine-grained cognitive parameters.

flowchart TD
    L1["Level 1: Foundation & Modes\n(Direct Mode · Userfield Search · Socratic Sparring · Zen Viewport)"] --> L2["Level 2: Context & Export\n(File Transfer · Media Pool · A4 Decks · Quick-Fire QR)"]
    L2 --> L3["Level 3: Specialized Studios & Labs\n(Director's Studio · Spatial Canvas · Focus Lab · Deep Research · 3D CAD · Visual Science Lab)"]
    L3 --> L4["Level 4: Architecture & Controls\n(Workflow Architect · EU Privacy Gateway · Model Router · The Critic)"]

1. Level 1: Foundation & Thinking Modes

Every session in the Brainstorm Reactor begins in one of two fundamental operating environments.

1.1 Direct Mode vs. Workflow Pipelines

• Direct Mode:
  Engineered for rapid, iterative synthesis. The operator supplies an intent, and the engine executes a single, highly disciplined cognitive transformation without pipeline overhead.
• Workflow Pipelines:
  Multi-agent execution structured according to the classic 3-Act thinking standard: Context Distillation → Dialectical Red-Teaming → Final Strategic Synthesis. In Step-by-Step mode, the operator retains complete approval authority prior to advancing each stage.

1.2 Userfield Search & Discovery

The primary prompt dock is designed for frictionless keyboard navigation and instant method discovery:

• Inline Slash Menu (/):
  Typing / anywhere in the prompt bar summons an instant floating palette. Operators can search, filter, and switch across more than 200 codified mental models and recipes (such as TRIZ Contradiction Matrix, Inversion, or First-Principles Decomposition) without touching the mouse.
• Smart Assistant Shelf:
  Positioned immediately above the input dock, the shelf performs real-time intent triage as you type. It surfaces live method-matching chips with relevance scores, suggests matching workflow pipelines, offers a Golden Model Upgrade chip for complex cognitive tasks, and includes a Direct-to-Recipe Saver ([Save X methods as recipe]) to persist executed methods into a custom pipeline with a single click.
• Universal Command Palette (Cmd + K):
  The global keyboard command bridge. Instantaneously search and execute any method, workflow, studio tool, or theme setting from anywhere in the interface.
• Dynamic Ergonomic Placeholders:
  The prompt area cycles contextual hints (/ für Befehle, Rezepte & Workflows…, Methoden & Rezepte suchen… ⌘K, ⌘+V…) to keep keyboard shortcuts top-of-mind.

1.3 Socratic Sparring ([Gegenwind])

Standard large language models exhibit sycophancy: they reflexively flatter user inputs and agree with flawed premises. Socratic Sparring counteracts this failure mode by deploying three dialectical probing schools:

1. Elenctics: Isolates unproven assumptions and internal logical contradictions.
2. Maieutics: Surfaces implicit domain knowledge and latent tradeoffs.
3. Dialectics: Pits thesis against antithesis to forge an auditable synthesis.

Launch Socratic Sparring instantly via the [INTERVIEW] badge above the primary prompt area or via the Command Palette (Cmd + K).

1.4 Viewports & Cognitive Ergonomics

• 3-Zone Studio Cockpit: Left navigation deck (session history and workflows), central thinking viewport, and right inspector (catalog and Media Pool). See Core Interface & Pool.
• Zen Mode: Eliminates visual noise by collapsing all surrounding decks into a calm, centered reading line with a docked bottom input. Access via Darstellung & Fokus in the navigation header.
• 50:50 Splitscreen: Automatically divides screen real estate during sparring or interview sessions, preserving the original briefing on the left while displaying adversarial counter-probes on the right.

1.5 The Mentor (Universal Cognitive Navigator)

The Mentor is an omnipresent guidance system accessible across Direct, Workflow, and Zen modes without switching viewports or disrupting your active working state:

• Global Shortcut (⌥M / ⇧⌘M): Instantly toggles the inline mentor dialogue in the active chat canvas.
• Top Navigation Bar ([ 🧭 Mentor ]): Centrally anchored in the top header with a breathing state beacon when active.
• Prompt Dock ([ 🧭 Mentor ]): Integrated directly alongside the model selector and search toggles in the prompt toolbar.
• Left Sidebar: Anchored under Interview & Diskurs.
• Dismiss & Clean Chat (Esc or [ESC ✕]): Instantly dismisses the active Mentor dialogue and purges all mentor notes from the chat canvas to restore an uncluttered workspace.
• Pause / Resume (⏸ Pausieren / ▶ Fortsetzen): Temporarily pauses or resumes the Mentor without purging dialogue history.

The Mentor conducts a focused 2–3 turn dialogue to clarify domain scope, recommend fitting cognitive methods or recipes, switch operating modes, and highlight key interface features via non-disruptive Attentional Beacons (luminous focus halos).

2. Level 2: Context, Media Pool & Corporate Exports

High-stakes executive decisions require seamless interoperability between incoming source materials and formal presentation formats.

2.1 File Context Tray & In-Memory Transfer

Drop unstructured PDF, DOCX, or text files directly onto the input dock. Clicking [In Eingabe übertragen] strips proprietary Office formatting artifacts using browser-native AST parsers and stages the sanitized text directly into the prompt bar without permanent storage overhead. Read more in Inputs & Context.

2.2 The Media Pool & Team Pool

The Media Pool acts as an asynchronous holding buffer. Bookmark any output block using the bookmark icon on message cards or save snippets directly from the assistant kebab menu.

• Media & Notes Pool: Personal scratchpad to hold text excerpts, attachments, and concepts across recipe switches.
• Team Pool: Collaborative workspace partition where teams gather disparate thoughts and execute 1-Click Team Synthesis to distill fragmented ideas into a unified strategic memo. See Pool Distribution.

2.3 Publication-Grade Multi-Format Export

• A4 Landscape Concept Deck (PDF): Optimized for 16:9 and 16:10 widescreen projection, client presentations, and C-level executive summaries. See Session Export.
• A4 Portrait Strategic Dossier (PDF): Formatted with book-printing typography and precise margins for formal memorandums and technical filings.
• Structured Data & Archives (ZIP, Markdown, HTML, CSV, DOCX, JSON): Complete session archives and table exports for downstream integration into Notion, Obsidian, Word, or internal data pipelines.

2.4 Session Sharing & Quick-Fire QR

• Session Sharing (/share/[id]): Generates an ephemeral, read-only web view featuring an upfront Executive Verdict. External stakeholders can inspect results and submit structured feedback (Approve, Request Pivot, Reject) without creating an account. Includes frictionless P2P sharing via WhatsApp, Email, or Web Share API, protected by a Cognitive Gate.
• Quick-Fire QR Code: Accessible via the input toolbar [+] menu (QR-Code erstellen) or Cmd + K. Instantly generates verified ISO/IEC 18004 compliant QR codes (downloadable as SVG or PNG) directly from prompt URLs or session share links.

2.5 Assistant Message Actions (Kebab Menu Suite)

Every assistant response features an unobtrusive action bar at the bottom left:

• One-Click Clipboard Copy: Instant markdown copy with visual checkmark confirmation.
• Run as Prompt: Injects the completed assistant synthesis back into the prompt dock for iterative refinement.
• Branch into New Session: Spawns an isolated session fork carrying the selected message as permanent context.
• Pin to Context: Pins critical findings permanently to the session context tray.
• Save to Pool: Routes text or selection into the Media or Team Pool.
• Attentional Beacons: The Inline Navigator dynamically illuminates these exact triggers with an animated, non-disruptive halo during guidance dialogues.

3. Level 3: Specialized Studios & Labs

For multimodal, spatial, and analytical tasks, the Brainstorm Reactor provides dedicated operational consoles, accessible from the unified Studios & Spezial-Formate ([ ⊞ Studios ]) dropdown in the top navigation bar:

• Briefing-Interview & Gegenwind Sparring: Rapid conversational scope discovery and dialectical thesis sharpening.
• Ideen-Labor & Director's Studio: Tactile cinematographic optics and infinite spatial node topologies.
• Visual Science Lab & Diagramm-Studio: Visual asset auditing and architectural flowchart synthesis.
• Simulation Cockpit (Alt + A · Admin): Live presentation orchestration and demo controls.

3.1 Director's Studio (/studio)

The Director's Studio replaces conversational prompt tinkering with tactile cinematographic controls:

• Camera & Optics: Direct calibration of physical focal lengths (24mm wide, 50mm natural, 85mm portrait), aperture settings (f/1.4 to f/11), film stocks, and lighting geometries (Chiaroscuro, Rembrandt).
• Style Lenses: Curated architectural, cinematic, and materials shaders.
• Prompt Ribbon Syntax: Binds visual parameters into a coherent natural language sentence that synchronizes bidirectionally with the underlying engine.

3.2 Idea Lab & Spatial Canvas (/lab)

An infinite visual thinking workspace that converts linear chat outputs into spatial node topologies. Features automated Dagre layout generation for dependency pipelines and semantic zooming to inspect both high-level macro clusters and granular sub-arguments. Learn more in the Idea Lab Guide.

3.3 Focus Lab (Neuro-Acoustic Instrument)

Accessible via the headphone icon in the navigation bar, Focus Lab provides a browser-native WebAudio synthesizer with 5 generative noise and resonance layers. It operates completely client-side without API latency to eliminate distractions during high-demand analytical work.

3.4 Deep Research Agent

For comprehensive market, patent, and technological investigations, the Deep Research Agent executes autonomous multi-step research loops. It queries authoritative sources, verifies factual claims with enforced citations, and compiles findings directly into A4 Strategic Dossiers. Launch via input toolbar [+] → Deep Research or Cmd + K.

3.5 3D Architekt & CAD Pipeline

A deterministic multi-agent pipeline generating validated Three.js code for physical geometries (2.5D profiles and volumetric bodies) complete with an interactive in-browser WebGL orbit viewer and export readiness for 3D printing. See Generative Tools & 3D Objects.

3.6 Visual Science Lab & Claim Audit (Admin Only)

An adversarial audit framework for evaluating creative concepts, ad visuals, and brand claims prior to public release. Uploaded campaign assets pass through 6 independent analytical perspectives (Concept, Touchpoints, Design VQI, Brand Compliance, Conversion, UI/UX) to isolate reception risks and bias. This suite requires administrative clearance (clearance: admin). See the Visual Science Lab Guide.

4. Level 4: System Architecture & Governance

Advanced users and enterprise teams maintain granular control over execution parameters and compliance protocols.

4.1 The Command Palette (Cmd + K)

The universal keyboard command bridge. Navigate directly to any method, launch the Socratic Navigator, switch between four calibrated dark mode variants (Neutral Pro, Slate, Warm Night, Pastel Contrast), or trigger document exports without leaving the keyboard.

4.2 EU Privacy Gateway & PII Redaction

Activating the green shield indicator enforces client-side and regional data privacy. Personally identifiable information (such as names, IBANs, phone numbers, and addresses) is redacted, and inference routes exclusively through certified European model endpoints. See EU Privacy Gateway.

4.3 Workflow Architect

Author, customize, and save multi-step thinking pipelines. Combine distinct reasoning operators into a proprietary company recipe, calibrate intermediate token allocations, and execute reproducible strategy frameworks. Learn more in Workflow Architect and Recipe Builder.

4.4 Cognitive Controls & The Critic

Fine-tune factual grounding, temperature corridors (0.8–1.2 SSOT), and model slot assignments in Experimental Settings. Attach The Critic as an adversarial verification layer to audit logic and citations prior to final rendering.

5. Diagnostic Tool Selection Matrix

Use this matrix to identify the ideal instrument for your current objective: