Skip to content

model-fallback.md

Latest commit

 

History

History
56 lines (44 loc) · 1.92 KB

model-fallback.md

File metadata and controls

56 lines (44 loc) · 1.92 KB

Model Fallback Chains

OpenClaw supports configuring multiple AI providers. Provider selection and fallback behavior are controlled entirely through environment variables - the operator's role is to inject the required API keys via envFrom or env.

How It Works

Fallback logic is application behavior - the operator does not intercept or manage LLM calls. The operator delivers the openclaw.json config and injects the required API keys.

When multiple provider API keys are present in the environment, OpenClaw can fall back between providers when a request fails (timeout, rate limit, 5xx).

Example CR

apiVersion: openclaw.rocks/v1alpha1
kind: OpenClawInstance
metadata:
  name: my-assistant
spec:
  envFrom:
    - secretRef:
        name: ai-provider-keys

The Secret ai-provider-keys should contain all required API keys:

apiVersion: v1
kind: Secret
metadata:
  name: ai-provider-keys
type: Opaque
stringData:
  ANTHROPIC_API_KEY: "sk-ant-..."
  OPENAI_API_KEY: "sk-..."
  GOOGLE_AI_API_KEY: "AIza..."

Required API Keys by Provider

Provider Environment Variable
Anthropic ANTHROPIC_API_KEY
OpenAI OPENAI_API_KEY
Google AI GOOGLE_AI_API_KEY
Azure OpenAI AZURE_OPENAI_API_KEY + AZURE_OPENAI_ENDPOINT
AWS Bedrock AWS_ACCESS_KEY_ID + AWS_SECRET_ACCESS_KEY
Mistral MISTRAL_API_KEY
Groq GROQ_API_KEY
DeepSeek DEEPSEEK_API_KEY
OpenRouter OPENROUTER_API_KEY

Notes

  • The operator webhook warns if no known provider API keys are detected in envFrom or env.
  • Each provider must have its API key available in the pod environment.
  • Rate limits and quotas are per-provider. A fallback chain spreads load across providers during outages but does not pool quotas.