ai.internal.make_cache

Construct a cache delimiter for structural prompt assembly. The `${cache(...)}` desugar in an LLM prompt body lowers to this call.

Reference version

Signature

function ai.internal.make_cache(args: baml.json.json = …) -> ai.CacheDelimiter throws never

Construct a cache delimiter for structural prompt assembly. The ${cache(...)} desugar in an LLM prompt body lowers to this call.

Source:<builtin>/ai/ns_internal/prompt.bamlbytes 2905–3032