Run LLM Prompt Batches on Render Workflows with OpenRouter Auto-Routing

OpenRouter ·

Key Info

Render Workflows can now run LLM prompt batches as individual task runs, with OpenRouter auto-routing each prompt and preserving the selected model in the result. You can scaffold a Node or Python project, develop locally, and deploy using Render's CLI.

Highlights

  • Start with render workflows init --language node and run a batch with render workflows start runPromptBatch.
  • Each prompt becomes its own queued, retried, and fanned-out task run on Render, so LLM batches don't need to live inside a web service.
  • OpenRouter's Auto Router (openrouter/auto) picks a model per prompt, and completion.model shows which model was used.
  • Local development uses render workflows dev; deployment uses render workflows create.
Loading...