Lab by Kraliki

Use Lab by Kraliki as a pre-configured multi-AI orchestration stack that delivers 16x productivity gains by combining specialized AI agents, semantic memory, and proven workflow patterns.

Architecture Overview

Review the architecture to understand agent roles, memory, and execution patterns.

┌─────────────────────────────────────────────┐
│  LAB BY KRALIKI                                  │
├─────────────────────────────────────────────┤
│  ORCHESTRATOR (Claude Opus)                 │
│  ├── Context & memory (mgrep semantic)      │
│  ├── Strategy & continuity                  │
│  └── Quality control                        │
│                                             │
│  WORKERS                                    │
│  ├── Gemini CLI (frontend, research)        │
│  └── Codex CLI (audit, backend)             │
│                                             │
│  PATTERNS                                   │
│  ├── Build → Audit → Fix                    │
│  ├── Parallel execution                     │
│  └── Hard problem voting                    │
│                                             │
│  MEMORY                                     │
│  ├── Semantic search (mgrep)                │
│  ├── Pattern library                        │
│  └── Decision docs                          │
└─────────────────────────────────────────────┘

Product Components

Use this table to track what is ready versus still in progress.

Business Model

Use this pricing table for positioning and sales conversations.

Margin: 80%+ (cost: €50-100/mo per VM)

Key Patterns

Apply these patterns when implementing Lab by Kraliki workflows.

Build → Audit → Fix

1. Gemini builds frontend/content
2. Codex audits for issues
3. Claude orchestrates fixes
4. Repeat until quality bar met

Parallel Execution

• Multiple workers on independent tasks
• Orchestrator manages context
• Results merged at checkpoints

Hard Problem Voting

• All agents propose solutions
• Best approach selected
• Execution with consensus

Folder Structure

Use this structure to locate product documentation, demos, and operational scripts.

magic-box/
├── README.md           # Full product concept
├── CLAUDE.md           # Project memory
├── SKILL.md            # This file (agentskills.io format)
├── docs/               # User documentation
│   ├── mvp-technical-spec.md
│   ├── onboarding-runbook.md
│   ├── cli-routing-guide.md
│   ├── beta-customers.md
│   └── landing-page-copy.md
├── templates/          # Customer CLAUDE.md templates
│   └── claude-md/      # By industry (agency, consultancy, legal)
├── demo/               # B2B demo materials
│   ├── scenarios/      # Demo workflows
│   ├── scripts/        # Demo automation
│   └── sample-projects/
├── stack/              # Docker/infrastructure configs
├── prompts/            # Polished prompt library
├── patterns/           # Workflow patterns
└── scripts/            # Setup and management scripts

Target Customers

Target these segments when building outreach and positioning.

Primary: Agencies & Consultancies

• Digital agencies needing capacity
• Consulting firms wanting leverage
• Freelancers wanting to scale
• Dev shops with backlog overflow

Value Prop: One person + Lab by Kraliki = 4-person team output

Secondary: Tech-Forward Companies

• Startups wanting speed
• Product teams with bottlenecks
• Innovation labs

Competitive Moat

Reference these advantages when drafting sales or marketing material.

Related Files

Development Commands

# Demo workflow
./demo/scripts/demo-start.sh

# Run specific scenario
./demo/scripts/run-scenario.sh agency-website

# Reset demo environment
./demo/scripts/demo-reset.sh

Lab by Kraliki: The team multiplier. Born December 2, 2025.