Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.wednesdayai.dev/llms.txt

Use this file to discover all available pages before exploring further.

OpenRouter

OpenRouter is a unified API gateway that gives WednesdayAI access to 200+ models from Anthropic, OpenAI, Google, Mistral, Meta, and many others — all from a single API key with a single billing account.

When to use OpenRouter

  • You want to experiment with many models without managing separate API keys
  • You need automatic fallback when a provider is down
  • You want unified billing across providers
  • You want access to models not available in your region

Authentication

Obtain an API key from openrouter.ai/keys.
openclaw login
# choose: OpenRouter
# enter your API key when prompted
Or set the environment variable:
export OPENROUTER_API_KEY=sk-or-v1-...
Lookup order:
SourceKey / path
Environment variableOPENROUTER_API_KEY
Stored credential~/.openclaw/credentials/openrouter

Selecting a model

OpenRouter model IDs use a provider/model format, prefixed with openrouter/:
{
  agents: {
    defaults: {
      model: "openrouter/anthropic/claude-opus-4-6",
    }
  }
}
Common model IDs via OpenRouter:
ModelWednesdayAI model ID
Claude Opus 4.6openrouter/anthropic/claude-opus-4-6
Claude Sonnet 4.6openrouter/anthropic/claude-sonnet-4-6
GPT-4oopenrouter/openai/gpt-4o
Gemini 2.0 Flashopenrouter/google/gemini-2.0-flash-001
Llama 3.3 70Bopenrouter/meta-llama/llama-3.3-70b-instruct
Mistral Largeopenrouter/mistralai/mistral-large
DeepSeek R1openrouter/deepseek/deepseek-r1
Run openclaw status --providers to list models your key can access.

Automatic fallback

OpenRouter supports provider fallback — if one provider is overloaded or down, it routes to another. This is configured in your OpenRouter account dashboard, not in WednesdayAI itself.

Troubleshooting

API key is invalid. Re-run openclaw login and enter a fresh key from openrouter.ai/keys.
Not all models are available to all account tiers. Check openrouter.ai/models to confirm your key has access to the model you selected.
OpenRouter adds a routing hop. If latency is a concern, consider connecting to the provider directly. Use OpenRouter for experimentation and fall back to direct connections for production.
OpenRouter may proxy to a different provider version than expected. Check openrouter.ai/models to see which backend a model ID resolves to.