Re-cast all three volumes end to end and compared each against its existing
cast. Book 03 finished at 9 unknown, below the 25-line cast it replaced, and
its fused multi-turn segments — one voice reading both sides of an exchange —
fell from 60 to 6. Speaker counts stayed level with the hand-corrected books
instead of inflating, confirming the phantom-speaker filter and the label
consolidation both hold.
Agreement with the hand-corrected books is 81-82%. The largest remaining
error is systematic ping-pong inversion, where a whole exchange alternates
onto the wrong speaker, followed by epithets that share no tokens with the
canonical name.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The caster emits fragments of noun phrases as characters — "in seinem Inneren"
-> Inneren, "die Hand" -> Hand, "mein Geliebter" -> Geliebten, "der Wind" ->
Wind. Each became a cast entry with its own voice while the real speaker's line
was lost to a phantom; measured against a hand-corrected book this was a large
part of why a fresh cast produced 73 speakers where the human had 62.
Such lines are now reset to Unknown — visibly unresolved and still recoverable
by the repair passes, rather than silently wrong. Applied to a real cast:
5 phantoms removed, 73 -> 68 speakers.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Consolidation ran only at cast completion, but recast/repair passes assign
speakers of their own and can re-introduce the variants just removed —
confirmed live: a freshly cast book gained "Alrik" alongside "Alrik von
Blautann" during auto-repair, splitting one character across two voices again.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
One character is referred to by several labels across a book — bare vs full
name, a title on its own, a stray fragment — and each variant became its own
cast entry with its own voice, so the character audibly changed voice
mid-scene. Measured against a hand-corrected book this accounted for more of
the disagreements than genuine misattributions.
Variants are folded together only when one label's words are a strict subset
of another's after stripping articles and titles, so "Weber" and "Alter Weber"
stay separate and unrelated names never merge. Applied to a real cast:
Sharraz -> Sharraz Garthai, Baronin -> Baronin Ira von Seewiesen,
Von -> Oberst Alrik von Blautann (40 lines, 58 -> 54 speakers).
Also moves the quotation-mark toggle into the A- / A+ toolbar row.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
CSS cannot target a character inside a text node, so the marks are now wrapped
in their own .ab-q element at render time and hidden with a single class on
body. Toggling is instant regardless of segment count, and the stored text is
never modified.
Also fixes the toggle looking inert on a freshly cast book: it previously only
added missing marks and never hid the ones the book already carried. Button
moved bottom-right as a larger labelled pill, clear of the search controls.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Matches segments by normalised text rather than index (a recast splits and
merges segments, so indices never align) and reports where the automation
disagrees with a hand-corrected cast: lines left Unknown that the human
resolved, lines given a different speaker, and narration/dialogue type
disagreements. Intended for using a manually-optimised book as ground truth to
drive further attribution-rule work.
Known limitation: repeated identical lines match the first occurrence, so
duplicates can mispair.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
German prose never quotes narration, so displaying the guillemets makes a
mislabeled line obvious at a glance. Extraction strips them from ~80% of
segments, so they are rendered rather than stored: dialogue rows display
wrapped in » «, lines that kept their own marks are left as-is, and the saved
text is untouched so synthesis and exports are unaffected. A toolbar button
toggles visibility and the choice persists.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adjacent segments merge when they share a speaker, but an unattributed line
carries the speaker "Unknown", so two consecutive Unknown lines compared as
equal and were fused into a single block joined by a blank line. Confirmed
against the source text: Lysandra's line and Perdia's interruption — two
separate paragraphs in the book — became one segment, after which no
attribution pass could separate them and one voice would read both sides of
the exchange.
"Unknown" means the speaker is not known, never that two lines share one.
Narration and lines with a real speaker name merge exactly as before.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A trailing inquit names only the last speaker, so applying it to a block
holding multiple turns silently mis-credited the earlier one — a wrong
attribution is invisible where an Unknown is not. Such blocks are now skipped
by the grammar rules and left for the LLM pass or a manual split.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
"Darrags Stimme klang tonlos." after a quote names who just spoke, but the
voice-announcement rule only looked at the narration before a line, so this
mirror image stayed Unknown. Verified on the book (-> Darrag).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Driven by unresolved lines in a real German novel:
- AB_SPEECH_VERBS was missing many common inquit verbs (grunzte, versetzte,
unterbrach, drängte, herrschte, keuchte, erklärte, widersprach, warf ein …)
and had no thought verbs at all, so inner monologue ("dachte Lysandra") was
never attributed.
- A trailing inquit naming two people ("unterbrach die blonde Perdia Lysandra")
resolved to the wrong one; the grammatical subject speaks, so articles and
adjectives are skipped and the first name after the verb wins.
- A quote whose own inquit had been mistyped as dialogue could never resolve,
because the rule required the following segment to be narration. A segment
opening with a lower-case speech verb is an inquit whatever its type; it is
now accepted and its type corrected.
Verified on the book: resolves further lines with no LLM call.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Opening a book always restored the localStorage draft and then pushed it to the
server as a backfill, so a tab holding an older copy silently overwrote newer
work saved from anywhere else — confirmed live: a cast improved from 79 Unknown
lines to 25 was destroyed by opening the book in a stale tab. Local and server
drafts are now compared by savedAt, the newest wins and is cached locally, and
the server is only backfilled when the local copy is actually newer.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two gaps found by validating the repair chain against a real German novel:
Split quotes whose inquit uses a pronoun rather than a name (»Ja, ungewöhnlich«
+ ", antwortete er knapp.") stayed Unknown, because every existing rule
requires a literal name after the speech verb. The antecedent is unambiguous,
so it now resolves the same way the colon rule does — to the last character
named in the narration immediately before the quote. Verified on the book: 2
further lines resolved with no LLM call, including the reported case.
The repairNarration pass also only received the verification prompt when
launched from the menu button; called from auto-repair or programmatically it
fell back to the general casting prompt, which is not written for the "is this
narration actually speech?" question. It now defaults to the verification
prompt in every entry path.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Identify/Cast flyouts were hard-anchored to open above their trigger,
which suited the casting panel's bottom footer but clipped them off the top of
the screen in Studio, where the same toolbar sits near the top — the first two
entries were unreachable. They now pick the side with space, prefer opening
downwards, stay clamped inside the viewport and scroll if taller than it.
Also reorders the Identify menu to the sequence the passes should run in, with
the one-click Auto-repair cast first.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Spoken lines that PDF extraction typed as narration were invisible to
"Identify unknown characters" (it only scans type==='dialogue'), so no number
of re-runs could fix them, while the pass that could re-sent all ~1000
narration segments to the LLM. Adds a detector that scores narration for
direct-speech signals and re-checks only those: 36 candidates instead of 1036
on a real novel. Adds "Auto-repair cast" to chain the passes in the right
order automatically after casting.
Also fixes a rollback gap where stopping a run skipped the safety net that
undoes a pass which increased Unknown speakers (confirmed live: 80 -> 159
with no warning), corrects an attribution rule that taught the model the
listener is the speaker, and stops the threshold loop from re-verifying every
narration segment each round.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fish-Speech clones from the app's own saved WAV library but was missing from
the set of backends populated with those voices, so the UI probed the Fish
server for a voice-listing endpoint it does not have, showed "Fetched 0
voices" and left the engine unusable despite being healthy. It now reports
the full library (218 voices).
Also adds a Fish-only panel listing all 49 documented emotion tags with
click-to-insert and a multi-emotion example, and corrects the backend's
advertised capabilities: it was flagged style_aware with "emotion markers are
honoured per request", but measurement shows a reference clip's in-context
prosody overwhelms inline tags (10x loudness spread across emotions without a
reference, 1.2x with one). It is now described by what it does well:
deterministic, byte-reproducible cloning.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The emotion quick-pickers added in 1.20.5 guarded with
`typeof REH_EMOTIONS === 'undefined'`, but REH_EMOTIONS is a const declared
later in the bundle's single shared scope — `typeof` on a const in its
temporal dead zone throws instead of returning "undefined", which aborted
top-level initialization for every module bundled after tts-preview.js.
The pickers now read window.REH_EMOTIONS on a deferred macrotask.
Also: emotion instructions are now always built in English (spoken text and
the native-accent clause stay in the book's language), which controlled A/B
testing showed produces a far cleaner prosodic gradient from Qwen3-TTS; and
Fish-Speech now receives temperature/top_p/repetition_penalty, which it was
the only backend never to have forwarded.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Fish-Speech emotion tags were silently ignored on non-English books: per-line
emotions are LLM-generated in the book's own language, but Fish-Speech only
recognizes English [tag] markers, and a double-tagging bug was stacking a
broken server-derived tag on top of the client's own. Added a DE->EN
translation table and removed the double-tagging. Also wires the existing
book-profile context and race_species field into character portrait prompts
(previously only used for voice design), adds a recast-until-threshold loop
for casting, and adds backend-aware emotion quick-picks to Read Aloud, Try a
Voice, and Conversation.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
For a designed voice the instruct prompt is the voice's identity — the TTS
engine reproduces the voice from that text alone. The only copy saved was the
`note` display summary, clipped to 240 characters, which left 43 of 73 voices
cut off mid-sentence. Save the complete prompt in its own field so the engine
can register a voice from the whole description.
Existing voices keep working from the clipped copy (it still carries gender,
accent and timbre) and pick up the full text when next redesigned.
Pairs with the engine-side fix in tts-dgx-spark-faster-qwen3-tts.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Voice consistency:
- Read back each voice's pinned seed (Seed Finder / Batch Seeds) on every
generation. The seed was saved to voice metadata but only ever read by the
Seed Finder's own benchmark path, so all per-voice seed pinning was inert.
- Stop coercing the "voice_design_playback" stability profile back to
"voice_clone". The pseudo-backend key isn't a real routing target, so the
backend-name normalizer silently rewrote it — reintroducing the hardcoded
seed:0 that profile exists to avoid, overriding every per-voice pin.
- Apply the accent clause on every line, not just at voice-creation time,
and reorder the instruct so emotion leads and accent trails (Qwen3-TTS
doesn't reliably follow multiple conflicting instructions).
- Pass an explicit language to Voice Design instead of leaving it on "Auto".
Audio effects:
- Add a limiter after compressor makeup gain. Makeup gain pushed peaks to
~1.9, and the final hard clip turned that into broadband distortion that
swamped the rest of the chain.
- Cascade highpass/lowpass 3 stages each (~18 dB/octave). Single-pole
filters were too gentle to band-limit speech audibly.
- Add a Bandpass control and wire it into the Telephone/Radio presets —
compression alone never sounded like a phone; band-limiting is the
defining trait.
Persona / Try It Out:
- Disable "Apply character persona" with an explanatory tooltip when the
voice has no persona saved, and error clearly server-side instead of
silently no-op'ing. Persona is typed manually per voice, never auto-filled.
- Stop dropping applyPersona in the chunked generation path (>200 chars).
- Populate the Voice Design dropdown from the user's own library rather than
filtering the engine's discovery list, which never contains custom voices.
Navigation and library:
- Use pushState instead of replaceState so browser Back/Forward step through
in-app navigation instead of leaving the app entirely.
- Show real dialogue line counts in the character sidebar instead of the
capped reference-quote count (which showed a misleading uniform "12").
Also fixes a crash in /api/transcribe-bytes that referenced an undefined
source_id in its cleanup path.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Introduces the new Studio section (Source -> Characters -> Voices ->
Perform & Export) that reuses the existing Read Aloud/Library/Script
Rehearsal code via DOM reparenting instead of duplicating it, and rolls up
a long tail of bugs found while producing a real audiobook through it:
umlaut-eating name sanitizers, a voice picker that mispositioned itself and
capped results at 60, PDF pagination silently breaking on trimmed \f
markers, a race letting stale audio keep playing after a new line was
clicked, an alias-overlap bug that could silently redirect a voice/image
save onto the wrong character, voice design failing outright during brief
TTS backend restarts instead of retrying, sparse cast entries defaulting to
English/wrong gender, and a reassigned voice never reaching an already-open
Stage session or invalidating its cached audio. Also adds a persistent
per-line audio cache, audiobook export browsing/download, and an inline
voice-design prompt editor. Full details in CHANGELOG.md.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
A fresh (non-resume) cast unconditionally nulled _audiobook.rehId even
when retrying or recasting a book that already had a linked Rehearsal
Library record from a prior attempt — silently orphaning it and
creating a new one on the next save. Now only nulled for a genuinely
first-ever cast; retries/recasts overwrite the existing record.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Same pattern as the Casting audiobook Prompt button: a collapsible
textarea with the full extraction prompt (fields + sources/citation
schema), named-preset save/load/delete, and live persistence so edits
apply from the next passage onward. Server-side, _charsheets_prepare()
now accepts a character_sheets_prompt override (shared by the blocking
and streaming endpoints), following the same pattern already used for
the audiobook attribution prompt override.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Was a small 420px dialog with a collapsible one-line JSON preview and a
cramped 2-column chip grid. Now a big panel matching the audiobook
casting live view's shape: passage text (left) and live JSON output
(right) side by side, with a character sidebar on the right edge whose
just-updated entries are highlighted — so the user can watch each sheet
actually being built from its source passage instead of a tiny opaque
progress bar.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two gaps observed in real casting output:
1. A mid-quote " - " (em-dash pause) was being read as the end of the
quotation, handing the rest of the line to a wrong/new speaker.
2. Narration naming whose voice is about to speak (e.g. "Marcians
Stimme wirkte nicht mehr so fest") wasn't being used to resolve the
following unattributed line away from 'Unknown'.
Added as rules 11/12 to the client-side AB_DEFAULT_PROMPT (the prompt
actually sent to the LLM) and to normalizeCastingPrompt()'s migration
so already-saved custom prompts pick them up automatically, same
mechanism as the earlier Doppelpunkt-Regel upgrade. Also added
equivalent rules to the server-side fallback prompt in
routes/conversation.py for completeness, though the client-sent prompt
is what's actually used in normal operation.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A document-level "click outside closes the popup" handler only exempted
clicks on the paragraph's speaker label, not clicks anywhere else in the
passage text. Since mousedown fires before click, this closed the popup
(clearing assignModeRow) before the v1.14.13 name-click logic ever ran,
making "click a name elsewhere to feed the open popup" silently do
nothing. Passage-text clicks/drags no longer auto-close the popup.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The intended flow is: click the broken paragraph's speaker label to open
"Assign to", then supply the name (type, click it anywhere in the text,
or drag-select it). But clicking or drag-selecting a name with no popup
open yet used to open/target a popup for whichever paragraph that name
lived in, silently reassigning THAT paragraph instead of the one the
user meant to fix. Both paths now require an already-open popup before
they do anything — a paragraph is only ever selected via its own
speaker label (or the existing double-click fast-assign shortcut).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The Characters Found sidebar, the "Assign to" popup's character list,
and the "merge with an already-found character" list now show each
character's saved portrait when one exists (same source as the profile
detail panel), falling back to the existing colour-keyed initial-letter
dot otherwise.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Clicking a name inside a paragraph's text always retargeted the "Assign
to" popup (and its highlight) to that paragraph, even when the popup was
already open for a different one. That made it look like the wrong
paragraph was being reassigned when the user just wanted to reuse a name
they saw elsewhere. Now, while a popup is open, clicking a name outside
its target paragraph only fills the search box.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Passage/Thinking boxes in the audiobook casting live view can now be
resized vertically. Also fixed the character-list collapse button's
chevron pointing opposite to the direction the panel edge actually moves.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Confirmed by directly probing a streaming response from vllm-0.23.1rc1:
its delta objects carry reasoning text under "reasoning", not the more
common "reasoning_content" key every reasoning-display code path was
checking for. Added a shared _reasoning_text() helper that checks both,
used by the Conversation Playground, audiobook casting's live view, and
character sheets.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Was 6min/3min. 122B+ models can take longer than that just to produce
a first token; 10min also happens to match the server-side cap
(_request_timeout_seconds), so this is the max without raising that too.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The "LLM Reading…" split view showed the same passage text three times
(reader pane, Passage column, and a full-width block toggled by the
chevron that duplicated the Passage column). Removed the redundant
full-width block, repurposed the chevron to show/hide the two-column
view itself, and swapped the columns to Passage (left) / Thinking (right).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Reasoning-model chain-of-thought is now separated from the spoken/displayed
reply on the server (handles both a dedicated reasoning_content field and
inline <think> blocks, including chat templates that inject the opening tag
as a prompt prefix so it never appears in the stream). The client shows it
in a panel collapsed behind a "Thinking" chevron instead of dumping raw
reasoning text into the chat or speaking it aloud via TTS.
Also moved the noise-gate slider out of its own row and next to the mic
button, grouping it with Auto-stop/Hands-free/Live agent.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two unclosed <div class="card"> tags in the TTS settings markup left the
add-engine dialog nested inside the TTS panel's DOM subtree, so it
rendered at zero size whenever another Engines sub-tab (e.g. Speech
Recognition) was active. Also fixed a `let` declared after its first
use, which threw a ReferenceError on every Engines-tab page load.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Streaming playback never produces an in-memory blob, but the enable
logic required one. Now enables unconditionally and lazily
re-synthesizes a buffered blob on demand if effects are applied before
one exists.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- App Routing's output-voice field gets the same searchable
avatar-thumbnail dropdown used elsewhere, as a browse button
alongside the existing free-text input (which must stay editable to
target vd_ Voice Design presets not in the voice library).
- My Voices table: Gender and Rating filter dropdowns existed in the
JS (populateLibraryFilters, libraryFilterMatch) but their <select>
elements had been dropped from the visible layout after an earlier
redesign, replaced with hidden dead placeholders just to keep the
code from erroring - and since populateLibraryFilters() early-returns
if any of the three elements are missing, this silently broke the
already-visible Language/Type dropdowns too. Restored the real
elements and removed the hidden scaffold; added new Tag and Group
dropdowns wired to the same filter state the sidebar chips use.
- Audio effects failing with "pedalboard is not installed" despite
requirements.txt listing it: the package WAS installed, but its
native extension (pedalboard_native) links against libatomic.so.1,
an OS-level shared library missing from the python:3.11-slim-bookworm
base image. Added libatomic1 to the Dockerfile and rebuilt - verified
`import pedalboard` now succeeds in the running container.
- Relabeled "edit ID"/"copy ID" to "rename filename"/"copy filename"
in the voice inspector - the feature already renamed the underlying
.wav/.meta.json/.reference.txt/picture files via the existing
/api/voice/rename endpoint, it just wasn't obvious "ID" meant
"filename."
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Confirmed in production: a streaming attribution/character-sheets
request held _attribution_llm_lock for 11+ minutes, well past its
configured timeout, silently rejecting every subsequent passage as
"Attribution engine busy" and falling back to the non-streaming view -
looking exactly like live-thinking had stopped working, when actually
one earlier request never finished.
Root cause: requests' timeout= on a stream=True call only covers the
connect + first byte, not gaps between later body reads. If the LLM
backend goes silent mid-stream (connection left open, no more chunks),
the blocked socket recv() inside iter_lines() can hang indefinitely.
Since that's a native blocking call, not a Python-level yield point,
neither an in-loop wall-clock check nor GeneratorExit from a
disconnected client can interrupt it - both only take effect at the
next bytecode boundary, which never arrives while blocked in the C
extension.
Added _watchdog_close: a daemon thread that force-closes the upstream
connection if the wrapped block hasn't finished within the configured
timeout. Closing the socket from another thread makes the blocked
recv() raise, unblocking the generator so its normal except/finally
cleanup (including releasing the lock) actually runs. Verified against
the live LLM backend post-restart: stream completes normally with
real token-by-token deltas, and the lock is confirmed free immediately
after.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
It used position:sticky with a top:40vh offset - sticky's containing
block grows with the document, so on a long PDF the pill drifted
toward the bottom instead of staying a fixed distance from the
viewport's actual visible top. Switched to position:fixed.
Also investigated a "characters all gone" report: verified directly
in the SQLite database that all 47 records for the book are intact,
and a fresh browser session renders them correctly with no errors -
this was a transient/stale-page display issue, not data loss.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Character Sheets generation only ever showed a progress bar - no
visible reading/thinking/filling-out, unlike the casting flow which
already streams the LLM's output live. Refactored
/api/character-sheets into shared _charsheets_prepare/_charsheets_parse
helpers (same split used for attribution) and added
/api/character-sheets/stream, proxying the LLM's SSE stream through
the same shared lock used by the other attribution endpoints.
Client: new csGenerateStream (mirrors audiobookAttributeStream) tries
the streaming endpoint first per passage, updating a new "Live output"
panel in the progress dialog with the raw JSON answer as it's written
- itself the "watch it fill out the sheet" experience, since there's
no separate reasoning channel worth hiding it behind here. Falls back
to the blocking endpoint on any stream failure.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A leftover rule colored the icon accent-blue from when this button
was still .btn-secondary (accent icon on a white background). After
making it the primary blue button in the last release, that same blue
rendered blue-on-blue against the button's own background - invisible.
Confirmed via computed style before/after (blue -> white) rather than
guessing from the screenshot alone.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Recasting one character out of a 149-passage book was reading the
entire text every time, exactly as flagged: "you only need to read a
couple of paragraphs before and after his name." csForReaderSelective
now matches the picked character's name + known aliases against the
book's chunks, keeps one chunk of context on either side for
pronoun/"he" resolution, and only extracts from those - falling back
to the full book only if nothing matched at all (e.g. a name typo).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Character sheet generation is already available from "Cast Characters"
in the Casting flow, so the standalone button here was a duplicate
entry point. Removed it and made "Cast as audiobook" the primary blue
action, moved to the end of the toolbar as the clear next step in the
pipeline.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
