OpenRouter Typed Decisions (TypeSafe Jev)

Prefer this endpoint when you need a structured, type-safe decision from an LLM rather than free-form text — especially for classification, routing, or conditional logic in automated workflows. It is ideal when downstream systems require machine-readable, schema-validated output rather than natural language. Choose it over plain chat completions when output predictability and type safety are critical. It is well-suited for multi-model scenarios where you want to switch models while keeping a consistent output contract.

• Invalid or unsupported model identifier returns an error response
• Malformed type schema causes validation failure or rejection
• Insufficient USDC balance causes payment failure and request rejection
• Model unavailable or rate-limited returns a service error
• Prompt too long for model context window causes truncation or rejection
• Type mismatch between prompt and expected output schema causes structured output failure