openova/products/axon

OpenOva Axon

SaaS LLM inference gateway connecting to OpenOva Cortex.

Status: Accepted | Updated: 2026-02-26

---

Overview

OpenOva Axon is a hosted inference gateway that provides managed access to LLM capabilities. It acts as the neural link between customer applications and OpenOva Cortex (self-hosted AI Hub), offering subscription-based LLM access without requiring customers to deploy their own GPU infrastructure.

flowchart LR
    subgraph Customer["Customer Environment"]
        App[Applications]
        Claude[Claude Code]
    end

    subgraph Axon["OpenOva Axon (SaaS)"]
        Gateway[API Gateway]
        Auth[Authentication]
        Metering[Usage Metering]
    end

    subgraph Cortex["OpenOva Cortex (Self-Hosted)"]
        KServe[KServe]
        vLLM[vLLM]
    end

    Customer --> Axon
    Axon --> Cortex

---

Key Features

• OpenAI-compatible API endpoint
• Subscription-based access with usage metering
• Automatic routing to optimal model instances
• Rate limiting and quota management
• Claude Code integration via ANTHROPIC_BASE_URL

Relationship to Cortex

Aspect	Cortex	Axon
Deployment	Self-hosted (customer cluster)	SaaS (OpenOva hosted)
GPU	Customer provides	OpenOva provides
Use case	Full AI platform	LLM inference only
Control	Full customization	Managed service

Usage

# Configure Claude Code with Axon
export ANTHROPIC_BASE_URL="https://axon.openova.io/v1"
export ANTHROPIC_API_KEY="your-subscription-token"

# Use Claude Code normally
claude "Explain this code..."

---

Deployment

Axon is a SaaS service operated by OpenOva. No customer deployment required.

For self-hosted inference, deploy OpenOva Cortex instead.

---

Part of OpenOva