[metadata]
apple-mobile-web-app-status-bar-style: default
apple-mobile-web-app-title: DuploCloud Documentation
color-scheme: light dark
generator: GitBook (05cff0a)
mobile-web-app-capable: yes
og:image: https://docs.duplocloud.com/docs/~gitbook/image?url=https%3A%2F%2F4059757114-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Forganizations%252FojpRPRrP7bqrzOUuLmOz%252Fsites%252Fsite_cLWUA%252Fsocialpreview%252FuXxSG8ZVhsF6IfNb38d2%252Fnew%2520purple%2520thumb.png%3Falt%3Dmedia%26token%3D6bcdbe8a-1bbf-4a07-9ef0-d0368bdc168b&width=1200&height=630&sign=7c210af52190be6153e47d7ad2b5ffe9&sv=3
og:title: Overview | DuploCloud Documentation
robots: index, follow
twitter:card: summary_large_image
twitter:image: https://docs.duplocloud.com/docs/~gitbook/image?url=https%3A%2F%2F4059757114-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Forganizations%252FojpRPRrP7bqrzOUuLmOz%252Fsites%252Fsite_cLWUA%252Fsocialpreview%252FuXxSG8ZVhsF6IfNb38d2%252Fnew%2520purple%2520thumb.png%3Falt%3Dmedia%26token%3D6bcdbe8a-1bbf-4a07-9ef0-d0368bdc168b&width=1200&height=630&sign=7c210af52190be6153e47d7ad2b5ffe9&sv=3
twitter:title: Overview | DuploCloud Documentation
viewport: width=device-width, initial-scale=1, viewport-fit=cover

[canonical-links]
https://docs.duplocloud.com/docs

[document-links]
/docs
AI Dashboards: /docs/armor/ai-dashboards
ARMOR Policy Model: /docs/armor/ai-devops-policy-model
AWS Extension: /docs/aws-extension/aws-extension
AWS User Guide: /docs/automation-platform/overview
Access Control: /docs/armor/access-control
Additional Components: /docs/armor/additional-components
Analytics: /docs/armor/analytics
Automation and Tools: /docs/automation-platform/automation-and-tools
Azure Extension: /docs/azure-extension/azure-extension
Azure User Guide: /docs/automation-platform/overview-2
Blog & News: https://duplocloud.com/blog/
Building an Ops Application: /docs/extension-framework/building-an-ops-app
CI/CD Extension: /docs/ci-cd-extension/ci-cd-extension
CI/CD Overview: /docs/automation-platform/introduction-to-ci-cd
Careers: https://duplocloud.com/careers/
Case Studies: /docs/extension-framework/case-studies
Compliance: https://duplocloud.com/solutions/security-and-compliance/
Contact: https://duplocloud.com/company/contact-us/
Container Orchestrators: /docs/automation-platform/container-orchestrators
Customer Stories: https://duplocloud.com/customer-stories/
Demo Videos: https://duplocloud.com/platform-demo/
DevOps Automation: https://duplocloud.com/solutions/devops-automation/
Developer Self Service via Slackbot: /docs/common-use-cases/developer-self-service-via-slackbot
Documentaiton: https://docs.duplocloud.com/
Duplo DevOps Agent: /docs/armor/agents
DuploCloud Prerequisites: /docs/automation-platform/duplocloud-prerequisites
Edge Deployments: https://duplocloud.com/solutions/deploy-on-the-edge/
Events: https://duplocloud.com/events/
Extras: /docs/automation-platform/extras-overview
FAQs: /docs/automation-platform/faq
FAQs: /docs/faqs
GCP Extension: /docs/gcp-extension/gcp-extension
GCP User Guide: /docs/automation-platform/overview-1
Getting Help with DuploCloud: /docs/getting-started/duplocloud-support-model
HelpDesk: /docs/armor/helpdesk
Home: https://duplocloud.com/
How Engineering and DevOps can work together: /docs/common-use-cases/how-engineering-and-devops-can-work-together
How Is Platform Engineering Evolving with AI — and What Is an Agentic IDP?: https://duplocloud.com/blog/agentic-internal-developer-platform
How to manage large complex projects: /docs/common-use-cases/how-to-manage-large-complex-projects
Incident Management: /docs/armor/incident-management
Installation: /docs/getting-started/installation
Integrating MCP Servers: /docs/getting-started/integrating-mcp-servers
Integrating Providers: /docs/getting-started/integrating-providers
Introduction: /docs/introduction
Kubernetes User Guide: /docs/automation-platform/kubernetes-overview
MCP Servers: /docs/armor/mcp-servers
Markdown: https://docs.duplocloud.com/docs/introduction/readme.md
Next The Three-Layer Architecture: /docs/introduction/the-three-layer-architecture
Observability Extension: /docs/observability-extension/observability-extension
Observability: /docs/automation-platform/diagnostics-overview
Onboarding: /docs/getting-started/duplocloud-onboarding
Overview: /docs
Overview: /docs/automation-platform/readme-1
Overview: /docs/extension-framework/extension-framework
Overview: https://duplocloud.com/platform/
Personas: /docs/armor/personas
Platform Engineering: https://duplocloud.com/solutions/devops-automation/
Platform: https://duplocloud.com/platform/
Plugin Architecture: /docs/extension-framework/plugin-architecture
Policy Model: /docs/automation-platform/application-focused-interface-duplocloud-architecture
Powered by GitBook: https://www.gitbook.com/?utm_source=content&utm_medium=trademark&utm_campaign=68cb0s9ce5UIUKWPuYs8&utm_content=site_cLWUA
Press: https://duplocloud.com/news/
Pricing Guide: https://duplocloud.com/pricing/
Pricing: https://duplocloud.com/pricing/
Privacy Policy: https://duplocloud.com/privacy-policy/
Product Updates: /docs/automation-platform/product-updates
Product Updates: /docs/product-updates
Projects: /docs/armor/projects
Providers: /docs/armor/providers
Public Cloud Tutorials: /docs/automation-platform/public-cloud-tutorials
Pulumi User Guide: /docs/automation-platform/pulumi-user-guide
Reporting AI Concerns: /docs/reporting-ai-concerns
Security Extension: /docs/security-extension/security-extension
Security and Compliance: /docs/automation-platform/security-and-compliance
Security and Compliance: /docs/common-use-cases/security-and-compliance
Settings: /docs/armor/settings
Skills: /docs/armor/skills
Studio: Developer Experience: /docs/introduction/studio-developer-experience
Support Options: Standard vs. Managed Operations: /docs/getting-started/duplocloud-support-options-standard-vs.-managed-operations
Terraform User Guide: /docs/automation-platform/terraform-support
The Extension Framework: /docs/introduction/the-extension-framework
The Three-Layer Architecture: /docs/introduction/the-three-layer-architecture
Tickets: /docs/armor/tickets
Trust Center: https://duplocloud.safebase.us/
User Administration: /docs/automation-platform/access-control
Webinars: https://duplocloud.com/webinars/
Workshops: /docs/automation-platform/workshops
Workspaces: /docs/armor/workspaces
https://www.gitbook.com/?utm_source=content&utm_medium=trademark&utm_campaign=68cb0s9ce5UIUKWPuYs8&utm_content=site_cLWUA
llms.txt: https://docs.duplocloud.com/docs/llms.txt

[content]
Overview | DuploCloud Documentation
⌘
Ctrl
k
Home
Platform
Pricing
Trust Center
More
Introduction
Overview
The Three-Layer Architecture
The Extension Framework
Studio: Developer Experience
ARMOR
ARMOR Policy Model
Providers
MCP Servers
Skills
Personas
Duplo DevOps Agent
Workspaces
Projects
HelpDesk
Tickets
Incident Management
AI Dashboards
Analytics
Access Control
Settings
Additional Components
FAQs
Product Updates
Reporting AI Concerns
Extension Framework
Overview
Building an Ops Application
Case Studies
Plugin Architecture
AWS Extension
AWS Extension
GCP Extension
GCP Extension
Azure Extension
Azure Extension
Observability Extension
Observability Extension
Security Extension
Security Extension
CI/CD Extension
CI/CD Extension
Getting Started
Installation
Integrating Providers
Integrating MCP Servers
Onboarding
Support Options: Standard vs. Managed Operations
Getting Help with DuploCloud
Common Use Cases
How to manage large complex projects
Security and Compliance
How Engineering and DevOps can work together
Developer Self Service via Slackbot
Automation Platform
Overview
Policy Model
Product Updates
Container Orchestrators
DuploCloud Prerequisites
Public Cloud Tutorials
AWS User Guide
GCP User Guide
Azure User Guide
Kubernetes User Guide
Security and Compliance
Terraform User Guide
Pulumi User Guide
Automation and Tools
CI/CD Overview
User Administration
Observability
FAQs
Workshops
Extras
Powered by GitBook
On this page
From DevOps to Platform Engineering — and What Changed
The Platform Engineering Problem, and What an Agent Needs
What Claude Code already solves — and what a platform adds
The enterprise maturity spectrum
Building an Agentic IDP
Detailed Implementation: The DuploCloud AI DevOps Platform
Was this helpful?
For the complete documentation index, see
llms.txt
. This page is also available as
Markdown
.
Copy
On this page
Introduction
Overview
From DevOps to Platform Engineering — and What Changed
Platform engineering exists to help DevOps scale: once every product team owns its own infrastructure, the combined weight of Kubernetes, Terraform, and compliance requirements is better handled with shared platform support than team-by-team. The industry's answer was centralization — stand up a platform team, build an Internal Developer Platform (IDP), encode best practices into "golden paths," and let developers self-serve instead of filing tickets.
In practice, that's proven harder to deliver than the pitch suggests. A real IDP — one with shared state, RBAC, audit trails, deterministic execution, and multi-engineer coordination — is a multi-year software build, and delivering one at that depth takes more time and engineering investment than most platform teams have had available. Many organizations stood up a platform team well before the platform itself existed. Golden paths often stayed aspirational, and ticket volume held steady.
AI's impact here has been concentrated in one area so far. It's shown up meaningfully in troubleshooting, largely through SRE-focused tooling, with more room to grow in deployments, release management, and compliance. The AI tool that got adopted at scale is the individual coding assistant — a huge productivity boost for one engineer at a time, and the next step is bringing that same boost to a whole team.
The shift that changes this: AI has made the economics of building a real IDP different. That's the idea behind an
Agentic IDP
— an Internal Developer Platform where every workflow is delivered as an agent, rather than as a static golden-path portal.
The Platform Engineering Problem, and What an Agent Needs
Look past the surface differences between one team's release process, another's environment setup, and a third's incident protocol, and most of that logic is the same everywhere — roughly 70% common functionality, 30% organization-specific.
An agent capable of covering this needs three things: a conversational interface, rich use-case-specific UX (forms, validations, dashboards, views — not everything should be a chat prompt), and the judgment to mix AI reasoning with plain, token-less API calls wherever a workflow doesn't actually need a model in the loop.
What Claude Code already solves — and what a platform adds
Engineers can already build a personalized agent with Claude Code that combines skills across Terraform, Jenkins, Kubernetes, and observability tooling. What a platform adds on top: a dedicated UI, APIs, token-less execution, and the multi-user, governance, and enterprise controls a platform team needs — since a personalized agent is architected as a single-user, single-session tool. That's the choice in front of most organizations: let individual engineers keep building their own agents against sensitive systems, or move toward a centralized platform with shared controls.
The enterprise maturity spectrum
Most organizations sit at one of three stages:
Small-scale copilots
— AI lives on individual laptops. Chat-only and single-player — great for one engineer, but without shared collaboration or enterprise controls yet.
Growing agentic
— Each use case builds its own multi-user plumbing independently: SSO, RBAC, access control, skill distribution, context management. Real capability, just not yet shared across use cases.
At-scale, centralized Agentic IDP
— A platform team consolidates the overlapping point solutions into one system. Very few organizations have reached this stage yet — it's where the rest are headed.
Building an Agentic IDP
The shape that solves this: a
common platform
that every use case builds on top of as its own custom application — plug-and-play, sandboxed from every other application, but inheriting the platform's capabilities automatically — paired with
a coding assistant specialized for building on that platform
(concretely, a Claude plugin) that enforces a standard implementation so applications stay interoperable instead of drifting into their own silos.
That gives two distinct personas: the
end user
, who uses an agent to get their actual work done, and the
agent builder
, who uses the specialized coding assistant to turn a new use case into a governed application on the shared platform.
Adapted from
How Is Platform Engineering Evolving with AI — and What Is an Agentic IDP?
— read the full essay for the complete argument.
Detailed Implementation: The DuploCloud AI DevOps Platform
This is exactly what the DuploCloud Platform implements, organized into three layers for building and running AI-native DevOps applications at scale.
At the foundation is
ARMOR (Agent Runtime for Multiplayer Operations)
— Layer 1. ARMOR is a Multiplayer Claude-as-a-Service runtime that brings together ticketing, connections, intelligence, workspaces, RBAC, projects, analytics, and cost management into a unified core. Business users — whether accessing through a web browser, REST API, Claude Code plugin, or Slack/Teams — interact with a shared Workspace that enforces role-based access control and scoped project context, while an underlying CaaS agent (built on the Agent SDK) handles ingestion, moderation, and communication with the LLM layer beneath it. This is what makes the platform
multiplayer
by design — the first of the three requirements above.
Building on that foundation, the
Extension Framework
(Layer 2) lets teams define policy models, build skills, and compose full Ops applications with forms, APIs, and lifecycle management — turning the runtime's capabilities into governed, repeatable workflows. This is the platform's answer to the second requirement,
custom workflows with rich UX
, and to the third — blending AI-driven steps with plain,
token-less
API execution wherever a model isn't needed.
On top of that sits the
Developer Experience & Hosting
layer (Layer 3), which allows teams to build and ship Ops applications in days using either a web interface or Claude Code, with flexible hosting options (managed or self-hosted). This is the
agent builder's
persona in practice — the specialized coding assistant for turning a new use case into a governed application on the shared platform, without building infrastructure from scratch.
Continue reading for a detailed walkthrough of each of the three layers — what they do, how they fit together, and what they unlock.
Next
The Three-Layer Architecture
Last updated
27 days ago
Was this helpful?
Platform
Overview
Demo Videos
Pricing Guide
Documentaiton
Solutions
DevOps Automation
Compliance
Platform Engineering
Edge Deployments
Resources
Blog & News
Customer Stories
Webinars
Privacy Policy
Company
Careers
Press
Events
Contact
© DuploCloud, Inc. All rights reserved. DuploCloud trademarks used herein are registered trademarks of DuploCloud and affiliates
