Longer mixed DE/EN phrase covering dotted numbers, compound nouns,
umlauts, special characters, English technical vocabulary, time
formats and prose — reveals more of a voice's character per seed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- voice-inspector: move `curGender` const before its first use — it was
declared at line 86 but referenced at line 61, causing a ReferenceError
(temporal dead zone) that silently aborted selectVoice() on every click
- Calc dB, Precompute, Batch Seeds now filter to checked voices when a
bulk selection is active, matching Benchmark's existing behaviour
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Auto-scroll to bottom is now gated on the user being within 80px of
the bottom. Scrolling up to review or edit a line freezes the feed
in place — new rows are still added, but the viewport doesn't move.
- Row-trimming is likewise suppressed while the user is scrolled up,
so old lines stay visible as long as they're being read/edited.
Trim limit raised from 80 → 600 rows so almost nothing is evicted.
- A floating "↓ Live" pill button appears at the bottom of the feed
whenever the user has scrolled up. Clicking it returns to the live
bottom and re-enables auto-scroll.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Casting view improvements:
- "LLM Reading…" row now shows a preview of the passage being processed.
A chevron button expands it to a full scrollable view of the passage
text (up to 260px), so you can follow what the LLM is reading live.
- Page-break dividers ("Page N") appear in the casting feed whenever the
source PDF page changes, giving a real-time view of page boundaries.
Bug fix:
- Page numbers in the Rehearser script were 0-indexed (PDF-internal) so
the first page break showed "Page 1" for what was PDF page 2, etc.
Now always emits 1-indexed page numbers (\f${page+1}).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
When a PDF audiobook is cast and saved to the Rehearser, page breaks now
carry the source PDF page number. In the Stage view each divider renders
as "— Page N —" instead of the generic "— Page break —".
Implementation:
- audiobook.js: `audiobookBuildScript()` encodes the page number in the
form-feed line (\fN instead of bare \f) at each page boundary.
- rehearser-parse.js: `parseScript()` now matches `line.startsWith('\f')`
and extracts the trailing page number into `line.page`.
- rehearser.js: `buildScriptPage()` renders "— Page N —" when `line.page`
is set; `toScript()` round-trips the number back (\fN) so it survives
save/reload; `importPDFScript()` also encodes page numbers (\f<pageIdx+1>)
when importing screenplay PDFs directly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Exposes the `speed` parameter (0.5–2.0) on the TTS `/v1/audio/speech`
request so audio is generated at the target tempo natively via the
faster-qwen3-tts backend rather than post-processing.
- Backend: `/api/tts-preview` now extracts and forwards a `speed`
override (clamped 0.1–4.0) through the same extra-params mechanism
already used for seed/temperature; backends that reject it fall back
cleanly via `_post_tts_with_fallback`.
- Try it out: "Native Speed" number input (0.5–2, step 0.05) added to
the text card; value persists in localStorage per browser; passed as
`extra` through `createTtsAudioSource` and `generateChunkedTts`.
- Read aloud: "Native Speed" control added to the generation controls
row alongside seed/temperature; included in `readerGenParams()` and
saved/restored with library documents (each book tracks its own speed).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
audiobookScopeText now records PDF page boundaries as character offsets in the
scope text; audiobookBuildScript realigns each segment against the source and
emits \f page-break markers at the nearest boundary. Saved/opened cast scripts
keep the book's pagination instead of collapsing to one flow.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Character sheets: stop sending the dead localhost:11434 default; fall back to
the server's configured llm_url so extraction uses the same working LLM.
- Audiobook casting: hoist highlightText to module scope so the "Review & cast"
manual-correction preview renders its segment rows again.
- Stage: unify narrator/dialog font size, add A-/A+ play text-size control
(scales A4 + paginated views), and make the cast chip list collapsible.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a persistent, book-scoped Character Library (new Characters section,
IndexedDB) that auto-fills from Character-sheet analysis with editable cards.
Enrich extraction with six narrative fields (backstory, relationships,
motivation, fears, mannerisms, voice/speech) plus the greyscale Good↔Evil
alignment bar, arc arrow, and 5-area Deep Analysis. Add ⋯ separators between
non-contiguous passages in Recast unknown, and harden dialogue attribution
against hallucination with same-language emotion tags.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Added a new 'Table View' button to the library sort bar.
- Implemented a CSS grid layout to display voice properties in a high-density table.
- Added sortable column headers (Name, Lang, Gender, Speed, dBFS, Length, Rating, Source, Seed, Note, Tags, Active).
- Aligned CSS grid to account for the bulk edit checkbox injection.
- Added drag-and-drop profile image support to the Inspector's large avatar.
- Ensured picture updates instantly synchronize across the List, Table, and Inspector views.
Read Aloud (new "Vorlesen" tab):
- PDF (real page render + overlay highlight) / TXT reader with live word
highlighting, voice + speed, per-sentence synthesis-state colours, zoom
(fit-width/height, two-page, ±), resume, and a server-side book library
(syncs across devices; per-unit MP3 audio fetched on demand).
Book -> multi-speaker audiobook:
- "Cast as audiobook" attributes dialogue to characters via the LLM
(guillemet/quote-style aware, turn-taking, recent-context), with a
deterministic speech-tag fallback. Editable preview, non-blocking live
casting panel, then auto-saved as a reopenable Script Rehearser play.
- Audiobook export: synthesise every cast line -> one MP3 per chapter.
Character sheets:
- LLM-extracted, self-filling RPG-style sheets (with page+quote sources)
in both Read Aloud and the Rehearser.
Also: MP3 storage + per-page/sentence export, voice-library "Precompute
embeddings" pre-warm, German "Vorlesen" i18n + flag language toggle,
large-PDF performance (lazy raster, buffer/canvas eviction, yielded parse),
and the Seed Finder changelog entry.
New: routes/reader.py, POST /api/attribute-dialogue, POST /api/character-sheets,
static/js/{reader,audiobook,character-sheets}.js, static/sections/s-reader.html.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds a collapsible "🎲 Seed Finder" panel to the voice inspector pane.
Select a voice, expand the panel, set a seed range (default 1–15), and
click Run to generate audio samples with different seeds.
Each result row has Play and "Use this seed" buttons. "Use this seed"
saves the chosen seed to the faster-qwen3-tts voices.json via the new
/api/tts-voice-seed proxy endpoint (which calls /voice-seed on the TTS
server). The saved row gets a ✓ marker.
Test sentence auto-selects DE/EN mixed or EN-only based on voice name.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
splitTextIntoChunks only split on sentence terminators (.!?), so structured
text separated by newlines was never chunked — the full page went as one
request, causing a TCP timeout that the browser reported as "Failed to fetch".
Fix: process each line individually before applying the sentence regex so
newline-separated segments become their own split points. Also move
generation.js from deferred batch E into the main feature batch C so
generateChunkedTts is always defined before the user can click Generate.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Storage: IndexedDB ('reh-library' DB, 'rehearsals' store).
Blobs (recorded mic audio) stored natively; no base64 overhead.
Library panel (Phase 1):
- List of saved rehearsals sorted by last update
- Each row: character-colour avatars, title, progress bar,
line count, recorded-clip count, date
- Active record highlighted in accent colour
- Open / Export .reh / Delete per-row buttons
- Import .reh button at the top
Transport bar (Phase 3):
- Save button (updates existing record if savedId set, else creates new)
- Export button (downloads current state as .reh JSON file)
Phase 4 (session complete):
- Save to library and Export .reh buttons
Data format (.reh file):
- JSON with version=1, title, script text, cast map, backend, lineIndex
- clips serialized with audio as base64 strings (mime + b64 fields)
- clipsFromJson restores Blob objects on import
Behaviour:
- Parsing a brand-new script resets savedId → safe to save as new record
- loadRecord() restores script, cast, lineIndex, clips and jumps to Phase 2
- renderLibraryList() called on init, after save, after delete, after import
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- A4 paper page (white, serif font, shadow) renders the full script at once
so you can read ahead while rehearsing
- Each dialog block shows: character avatar (voice library photo or initial
letter), character name in their colour, TTS/Me badge, and dialog text
- Active line gets a blue left-border highlight and auto-scrolls into view
- Transport bar (sticky, above the script):
- Cast strip: all character avatars at a glance
- ⏮ Prev / ▶ Play all / ⏹ Stop / ⏭ Next / 🔁 Repeat
- Progress bar + line counter
- Auto-play: TTS lines synthesize, play, auto-advance; 'me' lines pause
and slide up a sticky recording overlay at the bottom of the page
- Recording overlay shows the line to speak, oscilloscope + meter,
Record / Stop / Keep & continue / Re-record / Skip
- voice-library.js now exports window._voices so the rehearser can resolve
voice IDs to has_picture flags and fetch /api/voice/picture/{id}
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
## Script Rehearser (new feature)
- New section s-rehearser.html + rehearser.js + nav/loader wiring
- Phase 1: paste/upload script (.txt), auto-detect characters from
'CHARACTER: dialog' or ALL-CAPS screenplay format
- Phase 2: assign a TTS voice per character, or mark 'I play this'
- Phase 3: step-through rehearsal — synthesizes other characters via TTS,
shows level-meter + oscilloscope for your own lines, records them from mic
- Phase 4: session summary with per-line audio playback + download
## Connect Apps
- Removed duplicate standalone MCP/speak/hotkey full-width cards
- Kept the integration-grid cards (they use the real server URL from JS)
- Added Global Hotkey Daemon as a proper integration card with snippet-hotkey
populated by integrations.js (uses proxyBase URL dynamically)
## About page
- GET /api/changelog endpoint reads CHANGELOG.md and returns it as text
- Collapsible 'Changelog' <details> card fetches and displays it lazily
## Try It Out
- Reorganised into three cards: Voice & backend / Text to synthesize / Generate
- Backend help panel moved below the voice row (not in the same flex row)
- Style instruction field gains a dynamic badge ('style-aware ✓' / 'weak style')
and a yellow warning when a non-style-aware backend is selected while the
field is filled — wired to both backend-select change and input events
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Loader:
- Post-init modules (engines, ai-backends, generation, conversation) now
load AFTER the skeleton is removed instead of before. The UI is visible
~500 ms sooner on average; those four modules load while the user is
already browsing Voices / Clone / Design.
- Removed the sttReady event approach that was triggering a duplicate
/api/stt-backends call; init.js already populates all STT selects once
on startup.
Skeleton:
- Replaced the card-grid placeholder with a two-column workbench skeleton
(voice list rows on the left + inspector placeholder on the right) that
matches the real My Voices layout.
Connect Apps:
- MCP Server, /speak REST endpoint, and Global hotkey daemon sections
moved from Settings → About to Connect Apps, where they belong.
- About page now has GitHub + Releases links instead.
Clone section:
- Added a hint note beneath the sample-text textarea reminding the user
to replace the placeholder name (Sam / Alex / Marco …) with their own.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Root cause: VERSION was not volume-mounted, so the server always reported
1.2.0, causing browsers to serve 1-year-immutable cached JS even after
code changes.
Fixes:
- docker-compose.yml: add ./VERSION:/app/VERSION:ro volume mount.
After `docker compose up -d`, the server reads the current VERSION file
and JS is cache-busted by the correct version string.
- loader.js: append session timestamp to _appVersion so every page load
generates a unique JS URL. JS is always fresh regardless of whether
VERSION is current, at the cost of one network round-trip per file per
session (acceptable for a local tool).
- s-clone.html: embed EN default text directly in the textarea so the
field is never empty even before JS runs.
- voice-clone.js: remove the 'skip if already filled' guard in
initCloneSampleText so navigating back always resets to the language
text; call refreshSttBackends on load with sttReady event fallback.
- stt.js: dispatch 'sttReady' event after all STT listeners are wired.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
1. Sample text: expose initCloneSampleText as a window function and call
it from nav.js runSideEffects when the clone section is activated,
ensuring the textarea is always populated even if the IIFE ran before
the element existed.
2. Better sample texts: all 8 languages rewritten to ~38 words / ~15 s,
first-person, phonetically rich, proper Unicode diacritics.
3. Live mic monitor: a level-meter (18-bar) + scrolling oscilloscope
canvas (ring-buffer, 300 px, colour-coded) added to the microphone
card. "Check level" / "Stop monitor" buttons start/stop it
independently; clicking Record starts it automatically.
Uses raw mic constraints (no echo-cancel / AGC) for cleaner voice clone
audio. Mic gain slider and dB readout included.
4. Recording quality: MediaRecorder now requests audioBitsPerSecond:256000
in both voice-clone.js and stt.js.
5. STT engine picker: Recognition engine <select> + Refresh button added
above the Auto-transcribe button in Step 3. refreshSttBackends() now
syncs both stt-tts-stt-backend and clone-stt-backend. The transcribe
call passes the chosen backend to /api/transcribe.
6. File input: explicit extension list added to accept= for OGG/OPUS.
7. CSS: .mic-live-wave style added (dark/light theme variants).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds the 'Read aloud' sample-sentence box back to the microphone card
in the Clone a Voice tab. Language switcher covers EN/DE/IT/ES/FR/PT/NL/PL
with phonetically diverse sentences (same as the My Voices panel).
Uses the existing .sample-read-box / .sample-sentence CSS so it looks
identical to the equivalent panel in the voice library.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sentence text in typing bubble:
- Backend sends 'text' field with each audio SSE event (the sentence
being synthesised)
- Frontend audio queue stores {url, text} pairs
- playNextAudio() writes the sentence text into the '...' typing bubble
when LLM tokens haven't arrived yet (convCurrentSentenceBubble)
- convCurrentSentenceBubble cleared as soon as first LLM token arrives
so normal streaming takes over seamlessly
VAD noise fixes:
- VAD_THRESHOLD: 0.01 → 0.02 (background noise no longer counts as speech)
- VAD_MIN_REC_MS: 400 → 800ms (8/10s wait before silence detection starts,
gives user time to begin speaking without initial noise triggering send)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Speech gate (silence detection):
- vadHadSpeech flag: VAD auto-stop cancels without calling STT when no
speech was detected (fixes "[STT] No speech detected → gibberish" loop)
- cancelNextBlob flag: onstop skips processBlob when VAD cancels silently
- vadLastVoiceMs: gates preview transcription on actual detected speech
(prevents "reich" hallucination on initial silence chunks)
Hallucination filter:
- Client: HALLUCINATION_RE strips known Whisper phantoms from preview
- Server: _is_hallucination() in generate() treats "reich" / "danke" /
"thank you" etc. as "No speech detected" → never reaches LLM
Latency:
- VAD_SILENCE_MS: 1500 → 1000 ms (sends 500 ms sooner per turn)
- VAD_MIN_REC_MS: 500 → 400 ms
- MediaRecorder timeslice: 2500 → 1500 ms (preview text updates faster)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Previously all 11 section HTMLs had to complete before JS loading began.
Now only the active section loads first (needed before JS runs so event
listeners find their DOM elements). The other 10 sections fetch in parallel
with utils + settings; an explicit await before the 9-module feature batch
guarantees all section HTML is injected before any feature module sets up
its DOM listeners. On a local server the HTML files finish well before
settings.js completes, so the await is typically instant.
Sequential steps cut from: sections(11∥) + utils + settings + features + ...
to: section(1) + max(utils+settings, sections(10∥)) + features + ...
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace serial LLM-wait-TTS with overlapped execution:
- LLM streams via background thread → asyncio.Queue (non-blocking event loop)
- _sentence_split() detects sentence boundaries in the token stream
- asyncio.create_task fires TTS for each sentence immediately — TTS for
sentence 1 runs while LLM is still generating sentences 2, 3, …
- Audio chunks stream to frontend in order as each task completes
- Time-to-first-audio drops from (LLM total + TTS total) to
roughly (LLM time-to-first-sentence + TTS latency for one sentence)
Frontend audio queue:
- enqueueAudio() / playNextAudio() chain multi-chunk responses seamlessly
- clearAudio() stops playback and cancels queue on new turn or mic click
- scheduleAutoMic() waits for queue to drain before restarting mic
- Error paths clear the queue to avoid stale audio playing after failure
Also fix missing contextlib import (silent bug when audio temp files
needed cleanup in the STT path).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Conversation playground:
- Live speech preview: MediaRecorder sends accumulated audio to
/api/transcribe-bytes every 2.5 s; interim Whisper result shown in
the text input field while recording. Web Speech API tried first as
a faster path when available (HTTPS/localhost).
- VAD auto-stop: AudioContext AnalyserNode measures RMS every frame;
auto-stops after 1.5 s silence with a visible countdown. Auto-stop
toggle to revert to click-to-stop.
- Hands-free mode: mic auto-restarts after the agent finishes speaking
via audio.ended event + generation-counter cancellation. Hands-free
toggle (on by default) to disable.
Navigation:
- Persist active section and sub-page in localStorage; hard-reload
returns to the same page instead of always jumping to My Voices.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
