What to Do When OpenRouter is Down?

Juliette Chevalier· Jul 21, 2025

When OpenRouter goes down, you loose control of your product's entire AI operation.

You can't fix the issue yourself. You can't switch providers or implement custom failover logic that quickly. You can only wait and hope the service comes back online quickly, while loosing your users' trust.

Table of Contents

• Set Up Your Own AI Gateway
• OpenRouter vs. Helicone AI Gateway
• How to Migrate from OpenRouter to Helicone AI Gateway
• The Bottom Line

Set Up Your Own AI Gateway

You no longer need to create your own custom routing logic just to run your AI gateway within your own infrastructure. In less than 5 minutes, you can set up a gateway that is cheaper (only pay for provider cost vs 5% markup), up to 7x faster, and fully-featured.

The Helicone AI Gateway is the lightest and fastest AI gateway on the market. Built in Rust, it gives you complete control over your AI infrastructure while maintaining all the benefits of multi-provider routing.

Fully open-source, customizable, and built for scale - you can self-host it on your own infrastructure or use our managed service.

OpenRouter vs. Helicone AI Gateway

How to Migrate from OpenRouter to Helicone AI Gateway

Migrating from OpenRouter to the Helicone AI Gateway is surprisingly straightforward. Here's a step-by-step migration guide.

Step 1: Set Up Your Provider Keys

Unlike OpenRouter where you buy credits, the Helicone AI Gateway uses your direct provider API keys.

# Add keys for the providers you use to your project's `.env` file.
export OPENAI_API_KEY="your-openai-key"
export ANTHROPIC_API_KEY="your-anthropic-key"
export GEMINI_API_KEY="your-gemini-key"

Cost Benefit: You'll pay providers directly with no markup, typically saving 5-10% compared to OpenRouter's credit system.

Step 2: Run the AI Gateway

Get started locally by running the following command.

npx @helicone/ai-gateway@latest

Your gateway is now running on http://localhost:8080.

To use in production, deploy the AI Gateway using Docker or any of our quick deployment options here.

Step 3: Configure Your Routing Strategy

Create an ai-gateway-config.yaml file and set up your routing strategy:

# Example setup - equivalent to OpenRouter's auto-routing
routers:
  production:
    # Always uptime & rate-limit aware, choose your prefered routing strategy
    load-balance:
      chat:
        strategy: latency  # Automatically picks fastest provider. Could also be weighted, custom, etc.
        providers:
          - openai
          - anthropic
          - gemini
    # Optional: Add caching for cost savings
    cache:
      directive: "max-age=3600"  # 1 hour cache
    # Optional: Add rate limiting for cost control
    rate-limit:
      per-api-key:
        capacity: 1000
        refill-frequency: 1m
    helicone: # Optional: for authentication & observability enabled
      api-key: "your-helicone-key"

You can find more examples & configuration options here.

Step 4: Update Your Application Code

Replace https://openrouter.ai/api/v1 with http://localhost:8080/router/production and update your model name.

import openai

openai.api_base = "http://localhost:8080/router/production"
openai.api_key = "api-key" # SDK requires a string, but all auth is handled by the gateway

response = openai.chat.completions.create(
    model="openai/gpt-4o-mini",  # 100+ models available
    messages=[{"role": "user", "content": "Hello, world!"}]
)

Step 5: Test Your Migration

curl -X POST http://localhost:8080/router/production/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello from my self-hosted gateway!"}]
  }'

The Bottom Line

The Helicone AI Gateway gives you all the benefits of OpenRouter's multi-provider routing while eliminating the risks of vendor dependency and reducing costs through intelligent caching, self-hosting, and connecting directly to providers. You get better performance, lower costs, and complete control over your AI infrastructure.

Your AI infrastructure is too important to leave in someone else's hands.

Ready to migrate?: Start with the Helicone AI Gateway Quickstart and have your self-hosted gateway running in under 5 minutes.

Need help? Join our Discord community where engineers share deployment strategies and troubleshooting tips.

Want to contribute? The Helicone AI Gateway is fully open-source. Check out the GitHub repository to contribute features or report issues.

The Helicone AI Gateway is fully open-source. Check out the GitHub repository to contribute features or report issues.