Skip to main content

πŸš€ Introduction to Stacktic


Note
This guide does not represent the full set of capabilities or guidelines for each service.
Detailed information is generated directly within the application stack repository, which contains the complete, up-to-date reference.
The purpose of this guide is to explain key concepts and provide examples of our framework, automation, and solutions.
For comprehensive details on every feature and service, please refer to the stack documentation.


Stacktic is a new concept in full-stack management - a logic framework that understands and automates relationships across your entire technology stack.

❗ The Problem​

Current Reality​

Companies are trapped between bad options:

Option 1: Hyperscalers & Managed Services

  • Chosen due to lack of in-house skills
  • False sense of "insurance"
  • Long-term commitment contracts
  • Lock-in with no portability
  • Pay premium for a "promise" of faster time-to-market

Option 2: Open-Source Kubernetes Solutions

  • Requires highly skilled teams or big budgets
  • Massive operational burden
  • Complex versioning, observability, data pipelines
  • Security concerns add more complexity

Hard Decisions Companies Face Daily​

  • Private container-based LLaMA vs fast-moving AI services (OpenAI, Gemini, Claude)?
  • Stay with managed services or migrate to containers/Kubernetes?
  • How to transition from VMs or Cloud Foundry to pure modrern app Kubernetes?

The impact of each choice is not only about saving millions in infa or Ops,it’s about meeting future demands for privacy, ensuring flexibility, and maintaining the ability to adapt and change.​

alt text

The Root Cause​

Relationships​

It's all about relationships:

  • Between services and databases
  • Between microservices
  • Between data pipelines and applications
  • Between RBAC, metrics, and observability
  • These are stack relationships

alt text


πŸ’‘ What Is Stacktic?​

Core: A Logic Framework​

Stacktic is fundamentally a logic framework that:

  • Integrates and understands relationships between all stack components
  • Automates these relationships - from application topology to ETL to Day-2 operations

alt text

The UI Layer​

On top of the logic framework, a modular platform engineering UI that enables:

  • Integration of anything (inside or outside Kubernetes)
  • Automation and connection of all components
  • Real version control for every element of the stack using metadata

Revolutionary: Stack Version Control​

Using metadata, Stacktic creates actual version control for your entire stack:

  • Migrate to new services or versions out-of-the-box
  • Duplicate stack versions for production, staging, QA, testing
  • Rollback an entire stack to a previous version
  • Redefine complexity as versioned, trackable changes

alt text


🌟 The Value​

Immediate Impact​

  • Day-0 planning: From months to hours
  • Deployment costs: 70%+ reduction
  • Ops overhead: Eliminated through automation (SRE, DataOps, SecOps)
  • Security: Automated, removing blockers
  • Engineers: Focus on customization and improvement, not configuration

Strategic Benefits​

  • Cloud-agnostic stacks: Single command deployment on any cloud
  • Infrastructure savings: 30-50% reduction
  • Data regulations: Full compliance support
  • Negotiation power: No vendor lock-in
  • Sovereignty: Private ownership and sovereignty principles fully supported

πŸ€– AI Governance, Automation & Control​

The Challenge: AI Meets Infrastructure​

AI agents are becoming the primary operators of infrastructure β€” diagnosing incidents, querying databases, reading logs, triggering deployments. But without structure, AI operates blind: raw kubectl access, no understanding of relationships, no boundaries.

Stacktic's Answer: Metadata-Driven AI Control​

Stacktic doesn't just generate infrastructure β€” it generates the metadata layer that governs AI:

  • Structured Metadata Feed: Every component, link, sub-component, and attribute in your stack is exposed as typed, queryable data through the Stack Agent API. AI never parses YAML or guesses service names β€” it queries structured JSON.

  • Auto-Generated MCP Server: When you draw links in Stacktic, an MCP (Model Context Protocol) server is generated per stack with typed tools for every connected service β€” databases, message queues, observability, deployments. Draw a link β†’ tools appear. Remove a link β†’ tools disappear.

  • Governance by Architecture: The topology you design IS the governance. What AI can see, access, query, and modify is determined by the links you draw β€” not by a separate policy layer that can drift.

  • Zero Raw Access: AI agents operate through scoped tools with variable substitution ({namespace}, {password}, {database}) β€” the Stack Agent resolves variables from metadata. AI never sees raw credentials or runs arbitrary kubectl commands.

  • Write-Access Gating: Every service connection has an independent write-access flag. Read operations are always available. Write operations (publish messages, trigger syncs, insert rows) only appear when explicitly enabled per service.

  • Multi-Stack Isolation: Each stack generates its own MCP with its own credentials and topology scope. Cross-stack boundaries are explicitly controlled via is_external flags.

What This Means​

Without StackticWith Stacktic
AI + kubectl = unlimited cluster accessAI + MCP = scoped tools, typed metadata
AI must guess relationships from labelsAI knows every link, direction, and dependency
Governance bolted on after deploymentGovernance auto-generated from topology
Manual policy maintenance and driftZero-drift β€” topology changes propagate instantly
Same AI access across all environmentsPer-stack isolation with independent credentials

The metadata you feed AI IS the control. Stacktic generates 360Β° stack metadata β€” components, links, sub-components, attributes, cross-stack boundaries β€” and feeds it to AI through governed tools. This isn't a feature. It's a fundamental shift in how AI operates infrastructure.


🏭 The Dark Open-Source Factory​

In manufacturing, a dark factory runs lights-out β€” fully automated, zero human intervention. Stacktic brings this model to open-source infrastructure:

An open-source dark factory for topology, relationships, and operations automation.

The Factory Pipeline​

StageWhat Happens
1. TopologyDesign your stack β€” components, links, sub-components, attributes
2. Generation70 open-source templates assemble production-ready Helm, K8s, secrets, Day-2 ops
3. IntelligenceStack Agent API extracts structured metadata β€” every relationship becomes queryable data
4. GovernanceMCP server auto-generated per stack β€” typed tools, credential scoping, write gating
5. AutonomousAI agents operate with full stack awareness β€” incidents, validation, diagnostics β€” lights out

Why "Dark Factory"​

  • Lights out: Once topology is designed and deployed, AI agents operate autonomously through governed metadata β€” no manual kubectl, no ad-hoc scripts
  • Open source: 70 community templates, no proprietary lock-in, fork and extend anything
  • Cloud agnostic: Same stack definition deploys to any Kubernetes on any cloud
  • Self-aware: The factory knows its own topology β€” every component, link, and dependency is metadata that feeds AI intelligence

🌟 Vision​

Stacktic bridges the gap between:

  • Disadvantages of managed services (lock-in, cost, limited control)
  • Disadvantages of pure open-source (time-to-market, operational overhead)
  • Advantages of both approaches

By bypassing skill gaps and reducing operational complexity, Stacktic delivers fully stable, automated, cloud-agnostic, and sovereignty-compliant full stacks β€” an open-source dark factory where AI governance is the architecture, not a layer on top.

Our vision: The dark open-source factory for full-stack automation β€” where topology drives generation, generation produces metadata, metadata governs AI, and AI operates infrastructure autonomously. No vendor lock-in. No raw access. No manual intervention. Just topology in, governed autonomous operations out.

alt text


Stacktic: The dark open-source factory β€” topology in, autonomous operations out.