diff --git a/.planning/phases/34-voice/34-CONTEXT.md b/.planning/phases/34-voice/34-CONTEXT.md
new file mode 100644
index 00000000..e56d4527
--- /dev/null
+++ b/.planning/phases/34-voice/34-CONTEXT.md
@@ -0,0 +1,41 @@
+# Phase 34: Voice - Context
+
+**Gathered:** 2026-04-03
+**Status:** Ready for planning
+**Mode:** Auto-generated (discuss skipped via workflow.skip_discuss)
+
+<domain>
+## Phase Boundary
+
+Users can speak to the assistant (Whisper STT) and hear responses read aloud (Piper TTS) — Piper pre-warms visibly so the first synthesis call does not appear broken, and voice is offered during onboarding based on hardware capability.
+
+</domain>
+
+<decisions>
+## Implementation Decisions
+
+### Claude's Discretion
+All implementation choices are at Claude's discretion.
+
+</decisions>
+
+<code_context>
+## Existing Code Insights
+
+Codebase context will be gathered during plan-phase research.
+
+</code_context>
+
+<specifics>
+## Specific Ideas
+
+No specific requirements — refer to ROADMAP phase description and success criteria.
+
+</specifics>
+
+<deferred>
+## Deferred Ideas
+
+None.
+
+</deferred>