OpenRouter
By OpenRouter
Access multiple LLM models through one API, using HTTP or an OpenAI-compatible SDK with provider selection and fallbacks for dialogue backends.
Start: Select a model slug and test a short chat completion from your backend before connecting the game. Keep API keys server-side, not in shipped clients.
Game workflow: Compare dialogue models with the same request shape. Configure provider order and allow_fallbacks; use require_parameters when a feature must be supported.
Limits: Game state, rules and output checks remain your application’s responsibility. Set suitable workspace/guardrail budgets and review provider-specific logging and retention policies. This review checked documentation only, not model calls or dialogue quality.
Documentation and sources: https://openrouter.ai/docs/quickstart · https://openrouter.ai/docs/guides/routing/provider-selection · https://openrouter.ai/docs/guides/privacy/provider-logging · https://openrouter.ai/docs/guides/best-practices/spend-controls