Oracle Diagram Generator

Oracle Diagram Generator - Complete Reference

Purpose

Generate professional, accurate OCI architecture diagrams using official Oracle icons and styling. Supports Draw.io XML templates, Python diagrams library, and Mermaid.js for various output formats and workflows.

Official Oracle Icon Library

Download Sources

• Primary: https://docs.oracle.com/en-us/iaas/Content/General/Reference/graphicsfordiagrams.htm
• Draw.io Library: Download XML and import via File → Open Library From
• GitHub: https://github.com/oracle-quickstart/oci-arch-template

---

Complete Draw.io OCI Shape Reference

Compute Services

mxgraph.oci.compute_vm           - Virtual Machine
mxgraph.oci.compute_bm           - Bare Metal
mxgraph.oci.container_engine     - OKE (Kubernetes)
mxgraph.oci.container_instance   - Container Instance
mxgraph.oci.functions            - Functions (Serverless)
mxgraph.oci.instance_pool        - Instance Pool
mxgraph.oci.autoscaling          - Autoscaling Configuration

Database Services

mxgraph.oci.autonomous_database  - Autonomous Database (ATP/ADW/AJD)
mxgraph.oci.database_system      - DB System (VM/BM)
mxgraph.oci.exadata              - Exadata Cloud Service
mxgraph.oci.mysql                - MySQL HeatWave
mxgraph.oci.nosql                - NoSQL Database

AI & Machine Learning

mxgraph.oci.generative_ai        - OCI Generative AI
mxgraph.oci.data_science         - Data Science Platform
mxgraph.oci.ai_vision            - AI Vision
mxgraph.oci.ai_language          - AI Language
mxgraph.oci.ai_speech            - AI Speech
mxgraph.oci.ai_document          - Document Understanding
mxgraph.oci.anomaly_detection    - Anomaly Detection

Storage Services

mxgraph.oci.object_storage       - Object Storage
mxgraph.oci.block_storage        - Block Volume
mxgraph.oci.file_storage         - File Storage (NFS)
mxgraph.oci.archive_storage      - Archive Storage

Networking Services

mxgraph.oci.vcn                  - Virtual Cloud Network
mxgraph.oci.subnet               - Subnet
mxgraph.oci.load_balancer        - Load Balancer (L7)
mxgraph.oci.network_load_balancer - Network Load Balancer (L4)
mxgraph.oci.api_gateway          - API Gateway
mxgraph.oci.fastconnect          - FastConnect
mxgraph.oci.drg                  - Dynamic Routing Gateway
mxgraph.oci.nat_gateway          - NAT Gateway
mxgraph.oci.internet_gateway     - Internet Gateway
mxgraph.oci.service_gateway      - Service Gateway
mxgraph.oci.waf                  - Web Application Firewall

Integration & Streaming

mxgraph.oci.streaming            - Streaming (Kafka)
mxgraph.oci.integration          - Integration Cloud
mxgraph.oci.events               - Events Service
mxgraph.oci.notifications        - Notifications
mxgraph.oci.queue                - Queue Service

Security & Identity

mxgraph.oci.iam                  - IAM
mxgraph.oci.vault                - Vault (KMS)
mxgraph.oci.bastion              - Bastion Service
mxgraph.oci.cloud_guard          - Cloud Guard
mxgraph.oci.data_safe            - Data Safe

DevOps & Management

mxgraph.oci.devops               - DevOps Service
mxgraph.oci.resource_manager     - Resource Manager (Terraform)
mxgraph.oci.logging              - Logging Service
mxgraph.oci.monitoring           - Monitoring
mxgraph.oci.apm                  - APM (Application Performance)

---

Draw.io XML Templates

Template 1: Basic Three-Tier Architecture

Template 2: AI RAG Platform

Template 3: Multi-Agent Factory

---

Python Diagrams Library

Installation

pip install diagrams
# Requires Graphviz: brew install graphviz (macOS) or apt install graphviz (Linux)

Complete RAG Platform Example

from diagrams import Diagram, Cluster, Edge
from diagrams.oci.compute import Container, Functions, OKE
from diagrams.oci.database import AutonomousDatabase
from diagrams.oci.network import LoadBalancer, ApiGateway
from diagrams.oci.storage import ObjectStorage
from diagrams.oci.connectivity import FastConnect
from diagrams.oci.devops import ResourceManager
from diagrams.generic.compute import Rack
from diagrams.onprem.client import Users

# Graph attributes for Oracle styling
graph_attr = {
    "fontsize": "20",
    "bgcolor": "#FFFFFF",
    "splines": "ortho",
    "nodesep": "0.8",
    "ranksep": "1.0",
    "fontname": "Arial",
    "fontcolor": "#312D2A",
    "pad": "0.5"
}

node_attr = {
    "fontsize": "11",
    "fontname": "Arial",
    "fontcolor": "#312D2A"
}

edge_attr = {
    "color": "#312D2A",
    "penwidth": "2.0"
}

with Diagram(
    "OCI Enterprise RAG Platform",
    show=False,
    filename="oci_rag_platform",
    direction="TB",
    graph_attr=graph_attr,
    node_attr=node_attr,
    edge_attr=edge_attr
):
    users = Users("Users")

    with Cluster("OCI Region - us-ashburn-1"):
        with Cluster("Public Subnet"):
            lb = LoadBalancer("Load Balancer\n$13/mo")
            apigw = ApiGateway("API Gateway")

        with Cluster("Private Subnet - Application"):
            with Cluster("OKE Cluster"):
                app = [
                    Container("Chat UI"),
                    Container("RAG API"),
                    Container("Admin")
                ]

            functions = Functions("Chunking\nFunctions")

        with Cluster("Private Subnet - AI"):
            with Cluster("OCI Generative AI"):
                embedding = Rack("Cohere Embed 4\n$0.0001/1K")
                llm = Rack("Cohere Cmd A\n$0.0015/1K")

        with Cluster("Private Subnet - Data"):
            adb = AutonomousDatabase("ADB 26ai\nVector Search\n$0 Free Tier")
            storage = ObjectStorage("Documents\n$0.0255/GB")

    # Data flow
    users >> Edge(label="HTTPS") >> lb >> apigw
    apigw >> app
    app[1] >> Edge(label="embed") >> embedding
    app[1] >> Edge(label="chat") >> llm
    app[1] >> Edge(label="search") >> adb
    functions >> embedding >> adb
    storage >> functions

Multi-Agent System Example

from diagrams import Diagram, Cluster, Edge
from diagrams.oci.compute import Functions
from diagrams.oci.database import AutonomousDatabase
from diagrams.oci.network import ApiGateway
from diagrams.oci.storage import ObjectStorage
from diagrams.generic.compute import Rack

with Diagram("OCI Multi-Agent Architecture", show=False, filename="oci_agents"):

    with Cluster("OCI Generative AI"):
        hub = Rack("Agent Hub")

        with Cluster("Supervisor Layer"):
            supervisor = Rack("Supervisor Agent\nCohere Command A")

        with Cluster("Specialist Pool"):
            agents = [
                Rack("Data Agent"),
                Rack("SQL Agent"),
                Rack("Code Agent"),
                Rack("Report Agent")
            ]

    with Cluster("Tool Layer"):
        tools = [
            Functions("Functions"),
            ApiGateway("REST APIs"),
            AutonomousDatabase("Select AI")
        ]

    with Cluster("Data Layer"):
        adb = AutonomousDatabase("ADB 26ai\n(Memory + Specs)")
        storage = ObjectStorage("Artifacts")

    # Orchestration flow
    hub >> supervisor
    supervisor >> agents
    for agent in agents:
        for tool in tools:
            agent >> Edge(style="dashed") >> tool
    tools[2] >> adb
    agents >> storage

NVIDIA NIM Deployment Example

from diagrams import Diagram, Cluster, Edge
from diagrams.oci.compute import BM, VM
from diagrams.oci.database import AutonomousDatabase
from diagrams.oci.network import LoadBalancer, Vcn
from diagrams.oci.storage import ObjectStorage, BlockStorage
from diagrams.oci.security import Vault
from diagrams.generic.compute import Rack
from diagrams.onprem.network import Internet

with Diagram("NVIDIA NIM on OCI", show=False, filename="oci_nim"):

    internet = Internet("Global Traffic")

    with Cluster("OCI - Multi-Region"):

        with Cluster("us-ashburn-1 (Primary)"):
            lb_us = LoadBalancer("Traffic Mgmt")

            with Cluster("GPU Cluster"):
                gpus_us = [
                    BM("BM.GPU.H100.8\n$30/hr"),
                    BM("BM.GPU.H100.8\n$30/hr")
                ]

            with Cluster("NIM Services"):
                nims_us = [
                    Rack("LLM NIM\nLlama 405B"),
                    Rack("Embed NIM"),
                    Rack("Guardrail NIM")
                ]

        with Cluster("eu-frankfurt-1 (EU Sovereign)"):
            lb_eu = LoadBalancer("Traffic Mgmt")
            gpu_eu = BM("BM.GPU.H100.8")

        with Cluster("Shared Services"):
            storage = ObjectStorage("Model Weights")
            vault = Vault("API Keys")
            adb = AutonomousDatabase("ADB 26ai")

    # Flow
    internet >> [lb_us, lb_eu]
    lb_us >> gpus_us
    for gpu in gpus_us:
        gpu >> nims_us
    nims_us >> adb
    storage >> gpus_us

---

Mermaid.js Templates

RAG Architecture

flowchart TB
    subgraph Sources["Data Sources"]
        S3[S3]
        SP[SharePoint]
        API[APIs]
    end

    subgraph Ingestion["OCI Ingestion"]
        STR[OCI Streaming]
        FN[FunctionsChunking]
    end

    subgraph AI["OCI AI Layer"]
        EMB[Cohere Embed 4$0.0001/1K]
        LLM[Cohere Command A$0.0015/1K]
    end

    subgraph Data["OCI Data Layer"]
        ADB[(ADB 26aiVector Search$0 Free)]
        OBJ[Object Storage]
    end

    subgraph App["Application"]
        CHAT[Chat UI]
        APIGW[API Gateway]
    end

    Sources --> STR
    STR --> FN
    FN --> EMB
    EMB --> ADB
    CHAT --> APIGW
    APIGW --> LLM
    LLM --> ADB
    OBJ --> FN

    style ADB fill:#C74634,color:#fff
    style LLM fill:#C74634,color:#fff
    style EMB fill:#C74634,color:#fff

Agent Hierarchy

flowchart TB
    HUB[Agent HubOCI Generative AI]
    SUP[Supervisor AgentCohere Command A]

    subgraph Specialists["Specialist Agents"]
        DA[Data Agent]
        SA[SQL Agent]
        CA[Code Agent]
        RA[Report Agent]
    end

    subgraph Tools["Tool Layer"]
        SEL[Select AI]
        REST[REST APIs]
        FN[Functions]
    end

    ADB[(ADB 26ai)]

    HUB --> SUP
    SUP --> DA & SA & CA & RA
    DA & SA & CA & RA -.-> SEL & REST & FN
    SEL --> ADB

    style HUB fill:#C74634,color:#fff
    style SUP fill:#C74634,color:#fff
    style ADB fill:#C74634,color:#fff

---

Generation Workflow

When generating diagrams:

1. Analyze Request

• Identify architecture pattern (RAG, Agent, Web, Analytics)
• List required OCI services
• Determine output format preference

1. Select Template

• Choose closest matching template
• Customize service icons and labels
• Add connections and data flows

1. Apply Oracle Styling

• Use Oracle Red (#C74634) for primary elements
• Use proper icon shapes from mxgraph.oci.*
• Add cost annotations where relevant

1. Generate Output

• For Draw.io: Complete XML file
• For Python: Executable script
• For Mermaid: Markdown-embeddable code

1. Validate

• Check all icon references are valid
• Verify connections make logical sense
• Ensure costs are from verified pricing

---

File Output

When outputting Draw.io XML:

• Save as .drawio or .xml extension
• User can open directly in Draw.io or diagrams.net
• File → Open → Select the XML file

When outputting Python:

• Save as .py file
• Run with python filename.py
• Outputs PNG image to same directory

Resources

Official Documentation:

• OCI Graphics for Diagrams
• OCI Architecture Center
• Draw.io OCI Library

Tools:

• Draw.io / diagrams.net
• Python Diagrams Library
• Mermaid.js

---

Generate professional OCI architecture diagrams with verified Oracle styling and icons.

---

Quality Checklist

Before finalizing diagrams:

Visual:

• [ ] Uses Oracle Red (#C74634) for OCI services
• [ ] Correct icon shapes (mxgraph.oci.*)
• [ ] Consistent spacing and alignment
• [ ] Clear connection lines with labels

Content:

• [ ] All services labeled clearly
• [ ] Cost estimates included (where relevant)
• [ ] Data flow direction indicated
• [ ] Security boundaries shown (VCN, subnets)

Accuracy:

• [ ] Services match actual architecture
• [ ] Connections are logically correct
• [ ] Region/AD placement accurate
• [ ] Pricing from current OCI pricing page

Presentation:

• [ ] Title clearly states architecture purpose
• [ ] Legend included for complex diagrams
• [ ] Exportable format (PNG, SVG, PDF)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: oci-ai-architects
• Source: oci-ai-architects/claude-code-oci-ai-architect-skills
• License: MIT
• Homepage: https://oci-ai-architects.github.io/claude-code-oci-ai-architect-skills

Install and usage instructions live in the source repository linked above.