Document voice integration and verified backup coverage

docs/RECOVERY.md

@@ -9,6 +9,14 @@
 - `/opt/mike-ai` einschließlich des bereitgestellten, teilweise uncommitteten Stacks,
 - die Volumes `router-images`, `router-state` und `portainer_data`.
 
+Der neue WebRTC-Sprachadapter liegt unter
+`/opt/mike-ai/stack/services/athena-realtime-voice`. Ein manueller Lauf der
+vorhandenen Backup-Software am 16. September 2026 um 15:02 Uhr hat ihn
+eingeschlossen: Im Archiv `athena-2026-09-16T13-02-56.tar.gz` wurden sowohl
+`compose.yaml` als auch `services/athena-realtime-voice/server.py` geprüft.
+Das OpenClaw-Plugin auf Unraid liegt außerhalb dieses Athena-Backups; seine
+Quelle ist im Git-Repository unter `integrations/openclaw-athena-talk` erfasst.
+
 Piper-Daten sind kein aktueller Sicherungsbestand. Modellgewichte unter
 `/data/models` und das reproduzierbare Whisper-Volume gehören nicht zu diesen
 Backup-Mounts. Ein Backup ausschließlich auf `/data` schützt nicht vor deren Ausfall.

integrations/openclaw-athena-talk/README.md

@@ -10,8 +10,10 @@ through the separate `services/athena-realtime-voice` service:
    tools,
 4. Athena Qwen3-TTS returns PCM audio to the Talk client.
 
-The older `gateway-relay` path synthesizes long replies incrementally. The
+The older `gateway-relay` path splits completed replies into short TTS clips.
-experimental WebRTC path currently synthesizes each agent reply as one clip.
+The WebRTC path streams Athena's PCM audio to the browser as it is synthesized.
+Both paths wait for OpenClaw's agent-consult result before TTS starts; neither
+speaks partial text while the agent is still writing.
 
 The provider intentionally uses half-duplex audio: microphone input is paused
 while a response is being transcribed, generated, synthesized, or played. This
@@ -84,12 +86,16 @@ proxies the SDP offer through the existing OpenClaw HTTPS origin. On
 synthetic spoken request traversed the real OpenClaw offer route, private
 WireGuard media connection, Athena Whisper, and Qwen3-TTS and returned audio.
 The installed browser code also recognizes the bridge's
-`openclaw_agent_consult` call. A real UI microphone-to-agent-to-speaker call
+`openclaw_agent_consult` call. A real Chrome microphone-to-agent-to-speaker
-has **not** yet been observed, so this remains experimental. The Mac setting
+call and a second question in the same session were observed on 2026-09-16.
+The first browser failures required the bridge to announce user and assistant
+transcript items with stable IDs and predecessor ordering. The Mac setting
 “Echtzeitweiterleitung über Gateway verwenden” should be off for the
 browser-owned WebRTC path. On the tested Mac, its UI switch returned to on
 after navigation even though the native preference was set to false; Mac UI
-behavior still needs verification. `gateway-relay` remains available as a
+behavior still needs verification. No OpenClaw core files were patched; the
+plugin lives in OpenClaw's persistent extensions directory and the provider
+is selected through `talk.realtime` configuration. `gateway-relay` remains available as a
 reversible fallback. Never
 select `provider-websocket` for this service. See
 `services/athena-realtime-voice/README.md` for setup and limits.