Retrieve a saved prompt

Retrieve saved prompt configuration for your application. Omit both selectors to resolve production. Supply either label or version, never both; reject duplicate, empty, or unknown parameters. Authenticate and authorize before cache validation. The next server request sees a completed label move, including rollback. Exact versions are immutable. Missing/unassigned labels never fall back. A successful 200 or authenticated 304 for an archived prompt records activity. Retrieval never executes a model.

curl -X GET "https://example.com/api/v1/prompts/string"

{  "id": "support-example",  "name": "Support assistant",  "versionId": "version-example-1",  "version": 1,  "messages": [    {      "role": "system",      "content": "Answer questions about the sample product."    }  ],  "inputType": "user_message",  "modelRouteId": "openrouter:openai/gpt-4o-mini",  "modelId": "gpt-4o-mini"}