AI-powered Platform Engineering
CAIPE is an open-source multi-agent system that automates platform operations — incident response, deployments, runbooks, and more — so your team can focus on building.
💡 Pronounced like cape 🦸 — just as a cape empowers a superhero, CAIPE empowers platform engineers with 🤖 agentic AI automation.
# Install CAIPE via setup script
$ bash <(curl -fsSL https://raw.githubusercontent.com/cnoe-io/
ai-platform-engineering/main/setup-caipe.sh)# Or via Helm
$ helm upgrade --install ai-platform-engineering \
oci://ghcr.io/cnoe-io/charts/ai-platform-engineering \
--version 0.5.9 -f your-values.yamlReal World Usage
Used in production
Teams building with CAIPE — from always-on AI assistants to enterprise-scale multi-agent platform automation.
How Splunk Built Forge on CAIPE
Splunk Cloud Platform team used CAIPE to build an always-on internal AI assistant, cutting engineer response times by 99% and automating support triage across 90+ channels.
Read more ↗
JARVIS: Multi-Agent System Design Deep Dive
A technical deep dive into how JARVIS — Cisco's internal platform AI assistant — is architected as a multi-agent system using CAIPE for superior performance and scalability.
Read more ↗
CAIPE Hands-On Workshop
Step-by-step labs covering single-agent design, multi-agent orchestration, RAG knowledge bases, and distributed tracing — learn by building.
Start learning →Our Mission
To redefine platform engineering by creating intelligent, secure, and scalable multi-agent systems that empower teams to focus on innovation, seamlessly manage complex infrastructures, and shape the future of cloud-native operations through the Internet of Agents.
Why CAIPE
Built for platform teams
From incident response to deployments — CAIPE turns manual platform ops into intelligent agentic workflows.
Multi-Agent Orchestration
10+ agents and MCP servers with a deterministic task builder.
Rich Web UI
Streaming chat, agent builder, task builder, and skills gateway.
Integrated Knowledge Bases
RAG + Graph RAG across ArgoCD, AWS, Jira, GitHub, Slack, and more.
Agent Memory
Cross-session fact extraction and per-user context persistence.
Enterprise Security
OIDC SSO, Okta RBAC, and policy-based tool restrictions.
Flexible Deployment
Helm, Docker Compose, Langfuse tracing — bring any OpenAI-compatible LLM.
Deterministic Workflows
Sequential agent pipelines with no hallucinated steps.
Custom Agents
Build agents with custom prompts, tools, and personas — no boilerplate.
BYO A2A & MCP
Plug in your own A2A agents or MCP servers as first-class participants.
Multi-Model Support
Claude, OpenAI, Gemini, or any compatible endpoint — swap without rewiring.
Multiple Clients
Web UI, Backstage plugin, Chat CLI, Slack Bot, and Webex Bot.
Pre-built agents for your platform stack — or bring your own A2A agent, MCP server, and build custom agents
Built by the platform engineering community, for the community
An open-source project built by platform engineers, for platform engineers. Governed under the CNOE Agentic AI SIG — a community focused on platform engineering through shared guidance, tooling, and IDP reference architectures.