Skip to content

Dodaios is an open-source agent OS for your desktop: it runs your coding agents, browser and apps by voice or chat, on your own machine. Formerly Personal Jarvis.

docs(adr): ADR-0034 states what 1.4.0 ships — realtime and text now, the classic pipeline half in flight

Harald080ea913e05d
diff --git a/docs/adr/0034-non-blocking-heavy-turns.md b/docs/adr/0034-non-blocking-heavy-turns.md
index 8f2026391f..c058551117 100644
--- a/docs/adr/0034-non-blocking-heavy-turns.md
+++ b/docs/adr/0034-non-blocking-heavy-turns.md
@@ -2,6 +2,12 @@
**Status:** Accepted · **Date:** 2026-08-18 · **Phase:** Voice UX + text surfaces (every engine, every provider)
+**Implementation status (v1.4.0):** the realtime engine (§2) and the text
+surfaces (§4) shipped in 1.4.0. The classic pipeline (§3) is in flight:
+`[voice].background_heavy_turns` is declared in the config and documented
+here, but its pipeline reader lands with that change — until then a heavy
+turn on the classic pipeline still waits inline, exactly as before.
+
## Context
A heavy turn — the tool model, a router round trip, a background agent —
diff --git a/jarvis/core/config.py b/jarvis/core/config.py
index aa1d5cfffd..38e1b0780a 100644
--- a/jarvis/core/config.py
+++ b/jarvis/core/config.py
@@ -2673,7 +2673,9 @@ class VoiceConfig(BaseModel):
# ordinary utterances are answered meanwhile, and the heavy answer is
# spoken at the next quiet turn boundary, tied back to the request. False
# = the pre-2026-08-18 inline wait (speech during the wait cancels and
- # recombines). Read per turn by the pipeline.
+ # recombines). Read per turn by the pipeline once the classic-pipeline
+ # half of ADR-0034 lands (in flight at 1.4.0 — the realtime half shipped
+ # first); until then the switch is declared and not yet read.
background_heavy_turns: bool = True
# Per-gap budget after which a stale pending fragment is silently
# discarded (user-mandated 2026-05-26 — was: flushed/spoken). NOT a total