Phases 2-6: detection, visual timeline, backend hand-off, voiceprints

Phase 2 (call detection): CallDetector using CoreAudio per-process mic
attribution (anarlog technique) — robust start+stop for Zoom/Teams/Signal/Meet,
ignoring our own recording; auto-record toggle. Built; pending live multi-app
confirmation by the user.

Phase 3 (visual timeline foundation): AppAdapter protocol + SpeakerObservation,
TimelineBuilder (hysteresis/overlap/self-merge/aliases), VisualTimeline (schema
1.1), TextRecognizer (Vision OCR), FrameSampler + GridCallAnalyzer (name OCR +
saturated-highlight active-speaker attribution), SignalAdapter, VisualObserver
(window capture; frames released, never saved; minimized->visual_gap, idle != gap).
Synthetic-frame tested; adapter geometry pending real Signal fixtures + live
VisualObserver validation.

Phase 5 (backend hand-off): SparkControlClient (multipart label-merge, sequential,
TLS-skip, 503 Retry-After/413), SessionPackager (chunk plan + WAV slice + timeline
slice/rebase), TranscriptAssembler + SpeakersFile, TranscriptPipeline. Validated
END-TO-END against the live backend (chunk -> label-merge -> speakers.json).

Phase 6 (voiceprints): VoiceprintStore (known_voiceprints, persist named
fingerprints, skip Unknown). Wired: 'Send to backend' button + transcript status,
auto-send toggle (default off) + self-name setting.

All adversarial-review findings fixed. App + XCTest suite build; tests pass.

Ten31Transcripts/Visual/SpeakerObservation.swift

@@ -0,0 +1,36 @@
+import Foundation
+import CoreGraphics
+import CoreVideo
+
+/// One per-frame observation from an app adapter: a participant tile, whether its
+/// active-speaker cue is showing, and where it is. `name` may be a full name,
+/// initials (Signal), or "" if unknown. `t` is seconds relative to the session t0.
+struct SpeakerObservation: Equatable {
+    let name: String
+    let speaking: Bool
+    let bbox: CGRect
+    let confidence: Double   // 0…1
+    let t: TimeInterval
+}
+
+/// Per-app screen-reading strategy. Each conferencing app gets one implementation
+/// that knows that app's tile layout, name placement, and active-speaker cue.
+/// Adapters must be testable offline against still-image fixtures.
+protocol AppAdapter {
+    static var bundleIDs: [String] { get }
+    var adapterVersion: String { get }
+    var preferredFPS: Int { get }
+
+    /// Analyze one frame; return the speakers visible and whether each is speaking.
+    /// Must process in-memory and never persist the frame.
+    func analyze(frame: CVPixelBuffer, at t: TimeInterval) -> [SpeakerObservation]
+
+    /// Optional: participant names from the app's Accessibility tree (Electron
+    /// apps like Signal expose these), preferred over OCR when available.
+    func namesFromAccessibility() -> [String]?
+}
+
+extension AppAdapter {
+    func namesFromAccessibility() -> [String]? { nil }
+    var preferredFPS: Int { 3 }
+}