offload: speech-to-text stays two stages (V-486)

The one-call audio path is refused by measurement, so the inventory says
so where a future caller would read it.

docs/offload.md

@@ -1,6 +1,6 @@
 # Offloading model work to the workstation
 
-*Last verified: 2026-08-03 @ 12530c8. Living doc: correct it in place, do not append.*
+*Last verified: 2026-08-05 @ b789676. Living doc: correct it in place, do not append.*
 
 Owner's call, 2026-08-02. Vikunja #483 is the umbrella. Tasks #484 to #487 are the
 work, and this file holds the shape and the rules all four must obey.
@@ -149,6 +149,13 @@ flips. It is wired anyway: `PhraseReminder` is on the same transport and is on.
 Then the embedder above, **whisper.cpp** in `mavsttd`, and **piper** in `mavttsd`.
 `mavwaked` uses no model at all: an energy-threshold VAD over 30ms frames.
 
+Speech-to-text stays two stages when it moves. One call carrying both a clip and the router
+prompt was measured on 05-08-2026. It scores 54.2% intent-only against 84.7% for whisper on
+homesrv, on the same 72 cases. The model transcribes clips it then routes wrong, so a long
+classification prompt and an audio part compete for attention. Transcribing on the
+workstation and routing the text scores 83.3% at p50 997ms. So the transfer buys 375ms and
+cleaner transcripts, not accuracy. See `docs/evals/2026-08-05-audio-in-routing.md`.
+
 ## Order
 
 1. **Transport** (#484). Nothing else is possible until a seam can cross a host.