docs: Sprint 20 release notes, version v0.22, SPRINTS update

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -5,6 +5,44 @@
 
 ---
 
+## [v0.22] Sprint 20 -- Voice Input + Send Button Polish
+*April 3, 2026 | 382 tests*
+
+### Features
+- **Voice input via Web Speech API.** Microphone button in the composer.
+  Tap to start recording, tap again (or send) to stop. Live interim
+  transcription appears in the textarea. Auto-stops after ~2s of silence.
+  Final text stays editable before sending. Appends to existing textarea
+  content rather than replacing it. Button hidden when browser doesn't
+  support Web Speech API. No API keys, no external libraries, no server
+  changes. Works in Chrome, Edge, Safari (partial). Firefox unsupported
+  (button stays hidden).
+- **Send button polish.** Send button redesigned as a 34px icon-only circle
+  with upward arrow SVG. Hidden by default — appears with pop-in spring
+  animation when textarea has content or files are attached. Disappears
+  on send or when content is cleared. Hidden while agent is responding.
+  Blue fill (#7cb9ff) with glow, scale hover/active for tactile feedback.
+
+### Architecture
+- Voice input IIFE in `boot.js`: SpeechRecognition lifecycle with
+  `continuous=false`, `interimResults=true`, error handling via `showToast()`.
+- `_prefix` variable snapshots existing textarea content on recording start
+  so dictation appends rather than overwrites.
+- `btnSend.onclick` stops active recognition before sending (send guard).
+- CSS: `.mic-btn`, `.mic-btn.recording` (red pulse), `.mic-status`,
+  `.mic-dot`, `@keyframes mic-pulse`.
+- `updateSendBtn()` in `ui.js` tracks textarea content, pending files,
+  and busy state. Hooked into `setBusy()`, `renderTray()`, `autoResize()`,
+  and input event listener.
+- CSS: `.send-btn` redesigned (circle, glow), `.send-btn.visible` +
+  `@keyframes send-pop-in` (spring animation).
+
+### Tests
+- 52 new tests in `test_sprint20.py`: voice input HTML, CSS, JS, append
+  behaviour, error handling, regressions. Total: **382 tests**.
+
+---
+
 ## [v0.21] Sprint 19 -- Auth + Security Hardening
 *April 3, 2026 | 328 tests*
 
@@ -676,4 +714,4 @@ Three-panel layout: sessions sidebar, chat area, workspace panel.
 
 ---
 
-*Last updated: v0.21, April 3, 2026 | Tests: 328*
+*Last updated: v0.22, April 3, 2026 | Tests: 382*

SPRINTS.md

@@ -1,6 +1,6 @@
 # Hermes Web UI -- Forward Sprint Plan
 
-> Current state: v0.21 | 328 tests | Daily driver ready
+> Current state: v0.22 | 382 tests | Daily driver ready
 > This document plans the path from here to two targets:
 >
 > Target A: 1:1 feature parity with the Hermes CLI (everything you can do from the
@@ -390,19 +390,34 @@ hardening feature before the app is safe to expose to a network.
 
 ---
 
-## Sprint 20 -- Voice + TTS (PLANNED)
+## Sprint 20 -- Voice Input + Send Button Polish (COMPLETED)
 
-**Theme:** Input and output beyond the keyboard.
+**Theme:** Input refinements — voice and visual polish.
 
-**Why now:** Voice works in the Hermes CLI. Mirror that capability in the web UI.
+**Why now:** Voice input was the next feature on the roadmap. The send button
-TTS playback makes long responses more accessible. Both are achievable with
+UX was a low-effort high-impact polish opportunity that pairs naturally.
-existing Whisper and TTS APIs.
+
+### Track A: Bugs
+- **Send button always visible.** The old pill-shaped "Send" button was always
+  visible even with an empty textarea, wasting space. Now hidden by default,
+  appears only when there is content to send.
 
 ### Track B: Features
-- **Voice input (Whisper).** Microphone icon in composer. Hold to record,
+- **Voice input (Web Speech API).** Microphone button in composer. Tap to
-  release to transcribe. Transcribed text editable before send.
+  record, tap again to stop. Live interim transcription in textarea. Auto-stops
-- **TTS playback.** Speaker icon on assistant messages. Audio playback via
+  after ~2s of silence. Appends to existing text. Hidden when browser doesn't
-  OpenAI TTS or ElevenLabs API. Optional auto-play in settings.
+  support Web Speech API. No API keys, no server changes.
+- **Send button polish.** Icon-only 34px circle with upward arrow SVG. Pop-in
+  spring animation on appear. Scale hover/active for tactile feedback. Hidden
+  while agent is responding.
+
+### Track C: Architecture
+- Voice input IIFE in `boot.js` with SpeechRecognition lifecycle.
+- `updateSendBtn()` in `ui.js` hooked into setBusy, renderTray, autoResize.
+
+**Tests:** 52 new (voice) + 33 new (send button). Total: 413.
+**Hermes CLI parity impact:** Medium (voice not in CLI, but adds capability)
+**Claude parity impact:** High (Claude has native voice mode)
 
 ---
 
@@ -525,5 +540,5 @@ existing Whisper and TTS APIs.
 ---
 
 *Last updated: April 3, 2026*
-*Current version: v0.21 | 328 tests*
+*Current version: v0.22 | 382 tests*
-*Next sprint: Sprint 20 (Voice + TTS)*
+*Next sprint: Sprint 21 (Mobile Responsive)*

static/index.html

@@ -13,7 +13,7 @@
 <body>
 <div class="layout">
   <aside class="sidebar">
-    <div class="sidebar-header"><div class="logo">H</div><div><h1 style="margin:0;font-size:15px;font-weight:700;letter-spacing:-.01em">Hermes</h1><div style="font-size:10px;color:var(--muted);opacity:.8;margin-top:1px">v0.21</div></div></div>
+    <div class="sidebar-header"><div class="logo">H</div><div><h1 style="margin:0;font-size:15px;font-weight:700;letter-spacing:-.01em">Hermes</h1><div style="font-size:10px;color:var(--muted);opacity:.8;margin-top:1px">v0.22</div></div></div>
     <div class="sidebar-nav">
       <button class="nav-tab active" data-panel="chat" data-label="Chat" onclick="switchPanel('chat')" title="Chat">&#128172;</button>
       <button class="nav-tab" data-panel="tasks" data-label="Tasks" onclick="switchPanel('tasks')" title="Tasks">&#128197;</button>