feat: Qwen CustomVoice engine + profile compatibility UX#373
feat: Qwen CustomVoice engine + profile compatibility UX#373
Conversation
… engine on selection - Show all voice profiles with unsupported ones grayed out (opacity) instead of hidden - Clicking a grayed-out profile selects it and auto-switches the engine to a compatible one - Sort supported profiles first, with info tip about compatibility at the bottom - Scroll to selected profile after engine/sort changes with safe margin - Fix engine desync on tab navigation by initializing form engine from store Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
📝 WalkthroughWalkthroughUpdates engine and profile synchronization across frontend and backend. Changes engine derivation logic to support both Changes
Sequence Diagram(s)(omitted) Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@app/src/components/Generation/FloatingGenerateBox.tsx`:
- Around line 128-137: The code casts backend engine strings to EngineValue and
calls form.setValue('engine', ...) without validating them, which can produce an
invalid form state; update the logic in the useEffect around selectedProfile to
check that the resolved engine (from selectedProfile.default_engine or
.preset_engine) is one of the allowed EngineValue members before calling
form.setValue('engine', ...), and if it is not valid either skip setting the
form or map/fallback to a safe default and optionally log/debug; additionally
extract the engine enum/allowed values used here into a shared constant (used by
generationSchema and EngineModelSelector) so validation is centralized and you
can perform a simple includes(...) check against that shared list before setting
the form value.
In `@app/src/components/VoiceProfiles/ProfileList.tsx`:
- Around line 21-34: The effect schedules a requestAnimationFrame and a
setTimeout without cancelling them, which can cause stale DOM writes after
unmount or when selections change; modify the useEffect that references
selectedProfileId/selectedEngine and cardRefs to capture the RAF id and timeout
id returned by requestAnimationFrame and setTimeout, and return a cleanup that
calls cancelAnimationFrame(rafId) and clearTimeout(timeoutId); additionally,
before mutating el.style inside the RAF callback, re-check that
cardRefs.current.get(selectedProfileId) still exists and that the component is
still mounted (or that selectedProfileId matches) to avoid applying
scrollMarginTop to stale elements.
- Around line 52-55: The isSupported predicate currently marks every non-preset
profile as supported when a non-preset engine is selected; update isSupported
(used with allProfiles and selectedEngine) so that for non-preset engines it
returns true only for profiles with voice_type !== 'preset' AND where
profile.default_engine is either unset/empty or equals selectedEngine; preserve
the existing preset branch that checks p.voice_type === 'preset' &&
p.preset_engine === selectedEngine so both preset_engine and default_engine
constraints are handled correctly.
In `@app/src/lib/hooks/useGenerationForm.ts`:
- Line 67: The code currently force-casts selectedEngine into
GenerationFormValues['engine'] when building the form default (engine:
(selectedEngine as GenerationFormValues['engine']) || 'qwen'); replace that
assertion with a runtime guard: validate selectedEngine against the allowed
engine values (e.g., check it exists in the GenerationFormValues engine
union/enum or in an allowedEngines array/Object.values) and only use it if
valid, otherwise fall back to 'qwen'; add or reuse a helper like
isValidEngine(selectedEngine) and reference it where engine is set to avoid
unsafe casts and protect downstream logic.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 812469e2-6744-4500-ae5e-27341763449f
📒 Files selected for processing (5)
app/src/components/Generation/FloatingGenerateBox.tsxapp/src/components/VoiceProfiles/ProfileCard.tsxapp/src/components/VoiceProfiles/ProfileList.tsxapp/src/lib/hooks/useGenerationForm.tsbackend/routes/profiles.py
| type EngineValue = 'qwen' | 'luxtts' | 'chatterbox' | 'chatterbox_turbo' | 'tada' | 'kokoro' | 'qwen_custom_voice'; | ||
| useEffect(() => { | ||
| if (selectedProfile?.language) { | ||
| form.setValue('language', selectedProfile.language as LanguageCode); | ||
| } | ||
| // Auto-switch engine if profile has a default | ||
| if (selectedProfile?.default_engine) { | ||
| form.setValue( | ||
| 'engine', | ||
| selectedProfile.default_engine as | ||
| | 'qwen' | ||
| | 'luxtts' | ||
| | 'chatterbox' | ||
| | 'chatterbox_turbo' | ||
| | 'tada' | ||
| | 'kokoro', | ||
| ); | ||
| // Auto-switch engine to match the profile | ||
| const engine = selectedProfile?.default_engine ?? selectedProfile?.preset_engine; | ||
| if (engine) { | ||
| form.setValue('engine', engine as EngineValue); | ||
| } else if (selectedProfile && selectedProfile.voice_type !== 'preset') { |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Inspect how engine metadata is typed and consumed
fd -i 'types.ts' --exec rg -n 'default_engine|preset_engine' {}
rg -n --type=ts --type=tsx -C2 '\b(default_engine|preset_engine)\b'Repository: jamiepine/voicebox
Length of output: 224
🏁 Script executed:
# Get the actual FloatingGenerateBox.tsx file content around lines 128-137
fd -i floatinggeneratebox -type f | head -5Repository: jamiepine/voicebox
Length of output: 232
🏁 Script executed:
# Find the types files that matched in the initial search
fd -i 'types.ts' -type fRepository: jamiepine/voicebox
Length of output: 232
🏁 Script executed:
# Search for EngineValue definitions
rg -n 'type EngineValue|EngineValue\s*='Repository: jamiepine/voicebox
Length of output: 237
🏁 Script executed:
# Understand form setup and validation
rg -n 'form\.setValue.*engine' -C 3 -iRepository: jamiepine/voicebox
Length of output: 4299
🏁 Script executed:
# Find GenerationFormValues definition
rg -n 'type GenerationFormValues|interface GenerationFormValues' -A 15Repository: jamiepine/voicebox
Length of output: 1512
🏁 Script executed:
# Look for form initialization and schema validation
rg -n 'const form = useForm|useForm.*GenerationFormValues' -B 2 -A 10Repository: jamiepine/voicebox
Length of output: 5975
🏁 Script executed:
# Check for zod/validation schema
rg -n 'generationSchema|engine.*z\.' -A 3Repository: jamiepine/voicebox
Length of output: 1016
🏁 Script executed:
# Get full FloatingGenerateBox.tsx context around engine assignment
sed -n '120,150p' app/src/components/Generation/FloatingGenerateBox.tsxRepository: jamiepine/voicebox
Length of output: 1346
🏁 Script executed:
# Get the full generationSchema definition
sed -n '15,33p' app/src/lib/hooks/useGenerationForm.tsRepository: jamiepine/voicebox
Length of output: 541
Add runtime validation for backend engine values before setting form state.
Lines 134–136 cast backend engine strings directly to EngineValue without checking validity. If the backend returns an unexpected value, form.setValue('engine', ...) accepts it and the form silently becomes invalid.
The form schema defines engine as a Zod enum (generationSchema), but form.setValue() doesn't trigger validation—only form submission does. This creates a window where the form state contains an invalid engine without error signaling.
Add a guard to validate against known engines:
Suggested fix
type EngineValue = 'qwen' | 'luxtts' | 'chatterbox' | 'chatterbox_turbo' | 'tada' | 'kokoro' | 'qwen_custom_voice';
+ const VALID_ENGINES = new Set<EngineValue>([
+ 'qwen',
+ 'luxtts',
+ 'chatterbox',
+ 'chatterbox_turbo',
+ 'tada',
+ 'kokoro',
+ 'qwen_custom_voice',
+ ]);
useEffect(() => {
if (selectedProfile?.language) {
form.setValue('language', selectedProfile.language as LanguageCode);
}
// Auto-switch engine to match the profile
const engine = selectedProfile?.default_engine ?? selectedProfile?.preset_engine;
- if (engine) {
+ if (engine && VALID_ENGINES.has(engine as EngineValue)) {
form.setValue('engine', engine as EngineValue);Also consider extracting the engine enum to a shared constant to avoid duplication across FloatingGenerateBox.tsx, generationSchema, and other components like EngineModelSelector.tsx.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@app/src/components/Generation/FloatingGenerateBox.tsx` around lines 128 -
137, The code casts backend engine strings to EngineValue and calls
form.setValue('engine', ...) without validating them, which can produce an
invalid form state; update the logic in the useEffect around selectedProfile to
check that the resolved engine (from selectedProfile.default_engine or
.preset_engine) is one of the allowed EngineValue members before calling
form.setValue('engine', ...), and if it is not valid either skip setting the
form or map/fallback to a safe default and optionally log/debug; additionally
extract the engine enum/allowed values used here into a shared constant (used by
generationSchema and EngineModelSelector) so validation is centralized and you
can perform a simple includes(...) check against that shared list before setting
the form value.
| useEffect(() => { | ||
| if (!selectedProfileId) return; | ||
| // Wait a frame for the DOM to update after re-sort | ||
| requestAnimationFrame(() => { | ||
| const el = cardRefs.current.get(selectedProfileId); | ||
| if (!el) return; | ||
|
|
||
| // Temporarily apply scroll-margin so it doesn't land flush at the top | ||
| el.style.scrollMarginTop = '180px'; | ||
| el.scrollIntoView({ behavior: 'smooth', block: 'nearest', inline: 'nearest' }); | ||
| // Clean up after scroll completes | ||
| setTimeout(() => { el.style.scrollMarginTop = ''; }, 500); | ||
| }); | ||
| }, [selectedProfileId, selectedEngine]); |
There was a problem hiding this comment.
Add cleanup for deferred scroll operations.
Lines 24-33 schedule requestAnimationFrame and setTimeout without cancellation. Rapid profile/engine changes can trigger stale scroll writes after unmount or subsequent selections.
Proposed fix
useEffect(() => {
if (!selectedProfileId) return;
- // Wait a frame for the DOM to update after re-sort
- requestAnimationFrame(() => {
+ let timeoutId: ReturnType<typeof setTimeout> | null = null;
+ const rafId = requestAnimationFrame(() => {
const el = cardRefs.current.get(selectedProfileId);
if (!el) return;
@@
- setTimeout(() => { el.style.scrollMarginTop = ''; }, 500);
+ timeoutId = setTimeout(() => {
+ el.style.scrollMarginTop = '';
+ }, 500);
});
+ return () => {
+ cancelAnimationFrame(rafId);
+ if (timeoutId) clearTimeout(timeoutId);
+ };
}, [selectedProfileId, selectedEngine]);📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| useEffect(() => { | |
| if (!selectedProfileId) return; | |
| // Wait a frame for the DOM to update after re-sort | |
| requestAnimationFrame(() => { | |
| const el = cardRefs.current.get(selectedProfileId); | |
| if (!el) return; | |
| // Temporarily apply scroll-margin so it doesn't land flush at the top | |
| el.style.scrollMarginTop = '180px'; | |
| el.scrollIntoView({ behavior: 'smooth', block: 'nearest', inline: 'nearest' }); | |
| // Clean up after scroll completes | |
| setTimeout(() => { el.style.scrollMarginTop = ''; }, 500); | |
| }); | |
| }, [selectedProfileId, selectedEngine]); | |
| useEffect(() => { | |
| if (!selectedProfileId) return; | |
| let timeoutId: ReturnType<typeof setTimeout> | null = null; | |
| const rafId = requestAnimationFrame(() => { | |
| const el = cardRefs.current.get(selectedProfileId); | |
| if (!el) return; | |
| // Temporarily apply scroll-margin so it doesn't land flush at the top | |
| el.style.scrollMarginTop = '180px'; | |
| el.scrollIntoView({ behavior: 'smooth', block: 'nearest', inline: 'nearest' }); | |
| // Clean up after scroll completes | |
| timeoutId = setTimeout(() => { | |
| el.style.scrollMarginTop = ''; | |
| }, 500); | |
| }); | |
| return () => { | |
| cancelAnimationFrame(rafId); | |
| if (timeoutId) clearTimeout(timeoutId); | |
| }; | |
| }, [selectedProfileId, selectedEngine]); |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@app/src/components/VoiceProfiles/ProfileList.tsx` around lines 21 - 34, The
effect schedules a requestAnimationFrame and a setTimeout without cancelling
them, which can cause stale DOM writes after unmount or when selections change;
modify the useEffect that references selectedProfileId/selectedEngine and
cardRefs to capture the RAF id and timeout id returned by requestAnimationFrame
and setTimeout, and return a cleanup that calls cancelAnimationFrame(rafId) and
clearTimeout(timeoutId); additionally, before mutating el.style inside the RAF
callback, re-check that cardRefs.current.get(selectedProfileId) still exists and
that the component is still mounted (or that selectedProfileId matches) to avoid
applying scrollMarginTop to stale elements.
| const isSupported = (p: (typeof allProfiles)[number]) => | ||
| isPresetEngine | ||
| ? p.voice_type === 'preset' && p.preset_engine === selectedEngine | ||
| : p.voice_type !== 'preset'; |
There was a problem hiding this comment.
isSupported should account for default_engine on non-preset profiles.
Lines 52-55 currently treat every non-preset profile as supported whenever a non-preset engine is selected. Profiles with a specific default_engine can be incorrectly marked as supported.
Proposed fix
const isSupported = (p: (typeof allProfiles)[number]) =>
- isPresetEngine
- ? p.voice_type === 'preset' && p.preset_engine === selectedEngine
- : p.voice_type !== 'preset';
+ isPresetEngine
+ ? p.voice_type === 'preset' && p.preset_engine === selectedEngine
+ : p.voice_type !== 'preset' &&
+ (!p.default_engine || p.default_engine === selectedEngine);📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| const isSupported = (p: (typeof allProfiles)[number]) => | |
| isPresetEngine | |
| ? p.voice_type === 'preset' && p.preset_engine === selectedEngine | |
| : p.voice_type !== 'preset'; | |
| const isSupported = (p: (typeof allProfiles)[number]) => | |
| isPresetEngine | |
| ? p.voice_type === 'preset' && p.preset_engine === selectedEngine | |
| : p.voice_type !== 'preset' && | |
| (!p.default_engine || p.default_engine === selectedEngine); |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@app/src/components/VoiceProfiles/ProfileList.tsx` around lines 52 - 55, The
isSupported predicate currently marks every non-preset profile as supported when
a non-preset engine is selected; update isSupported (used with allProfiles and
selectedEngine) so that for non-preset engines it returns true only for profiles
with voice_type !== 'preset' AND where profile.default_engine is either
unset/empty or equals selectedEngine; preserve the existing preset branch that
checks p.voice_type === 'preset' && p.preset_engine === selectedEngine so both
preset_engine and default_engine constraints are handled correctly.
| modelSize: '1.7B', | ||
| instruct: '', | ||
| engine: 'qwen', | ||
| engine: (selectedEngine as GenerationFormValues['engine']) || 'qwen', |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verify selectedEngine is not strongly constrained at store boundary
fd -i 'uiStore.ts' --exec sed -n '1,220p' {}
rg -n --type=ts --type=tsx -C2 '\bsetSelectedEngine\s*\('Repository: jamiepine/voicebox
Length of output: 2481
🏁 Script executed:
# Check useGenerationForm.ts around line 67 to confirm the code
fd -type f -name 'useGenerationForm.ts' -exec sed -n '40,75p' {} +Repository: jamiepine/voicebox
Length of output: 232
🏁 Script executed:
# Find GenerationFormValues type definition
rg -n 'type GenerationFormValues|interface GenerationFormValues' -A 15 --type tsRepository: jamiepine/voicebox
Length of output: 1512
🏁 Script executed:
# Find all usages of setSelectedEngine to see what values are being set
rg -n 'setSelectedEngine' --type ts -B 2 -A 2Repository: jamiepine/voicebox
Length of output: 1912
🏁 Script executed:
# Find generationSchema definition
rg -n 'generationSchema\s*=' -A 20 --type tsRepository: jamiepine/voicebox
Length of output: 1493
🏁 Script executed:
# Read FloatingGenerateBox around setSelectedEngine usage to see watchedEngine source
sed -n '110,135p' app/src/components/Generation/FloatingGenerateBox.tsxRepository: jamiepine/voicebox
Length of output: 1060
🏁 Script executed:
# Read useGenerationForm.ts around line 67
sed -n '60,75p' app/src/lib/hooks/useGenerationForm.tsRepository: jamiepine/voicebox
Length of output: 477
Guard selectedEngine at runtime instead of asserting it.
Line 67 force-casts a plain string from store into GenerationFormValues['engine']. The store accepts any string without validation, and while the current code path (form.watch → setSelectedEngine) maintains valid enum values, the store boundary is unguarded. If the store gets persisted, synced from external sources, or called from new code paths, invalid values can bypass form validation and break downstream logic.
Proposed fix
+const ALLOWED_ENGINES = new Set<NonNullable<GenerationFormValues['engine']>>([
+ 'qwen',
+ 'qwen_custom_voice',
+ 'luxtts',
+ 'chatterbox',
+ 'chatterbox_turbo',
+ 'tada',
+ 'kokoro',
+]);
+
+const initialEngine: NonNullable<GenerationFormValues['engine']> = ALLOWED_ENGINES.has(
+ selectedEngine as NonNullable<GenerationFormValues['engine']>,
+)
+ ? (selectedEngine as NonNullable<GenerationFormValues['engine']>)
+ : 'qwen';
+
const form = useForm<GenerationFormValues>({
resolver: zodResolver(generationSchema),
defaultValues: {
@@
- engine: (selectedEngine as GenerationFormValues['engine']) || 'qwen',
+ engine: initialEngine,
...options.defaultValues,
},
});🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@app/src/lib/hooks/useGenerationForm.ts` at line 67, The code currently
force-casts selectedEngine into GenerationFormValues['engine'] when building the
form default (engine: (selectedEngine as GenerationFormValues['engine']) ||
'qwen'); replace that assertion with a runtime guard: validate selectedEngine
against the allowed engine values (e.g., check it exists in the
GenerationFormValues engine union/enum or in an allowedEngines
array/Object.values) and only use it if valid, otherwise fall back to 'qwen';
add or reuse a helper like isValidEngine(selectedEngine) and reference it where
engine is set to avoid unsafe casts and protect downstream logic.
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
| isSelected && 'ring-2 ring-accent shadow-md', | ||
| 'cursor-pointer transition-all flex flex-col h-[162px]', | ||
| disabled ? 'opacity-40 hover:opacity-60' : 'hover:shadow-md', | ||
| isSelected && !disabled && 'ring-2 ring-accent shadow-md', |
There was a problem hiding this comment.
Disabled+selected card deselects instead of re-engaging
Medium Severity
When a profile is both disabled and isSelected (e.g., user selects a cloned profile then manually switches the engine to Kokoro), the selection ring is hidden via isSelected && !disabled, making the card look unselected. However, handleSelect still uses the toggle isSelected ? null : profile.id, so clicking it deselects the profile instead of re-selecting it and triggering the engine auto-switch. The FloatingGenerateBox's EngineModelSelector doesn't pass selectedProfile, so users can freely switch to an incompatible engine, hitting this state.
Additional Locations (1)
| } else if (selectedProfile && selectedProfile.voice_type !== 'preset') { | ||
| // Cloned/designed profile with no default — ensure a compatible (non-preset) engine | ||
| const currentEngine = form.getValues('engine'); | ||
| const presetEngines = new Set(['kokoro', 'qwen_custom_voice']); |
There was a problem hiding this comment.
Duplicated preset engine set across components
Low Severity
The set of preset engines (['kokoro', 'qwen_custom_voice']) is independently defined in both ProfileList.tsx (as PRESET_ENGINES) and inline inside a useEffect in FloatingGenerateBox.tsx (as presetEngines). A related complement set CLONING_ENGINES also exists in EngineModelSelector.tsx. Adding a new preset engine requires updating all three locations, creating a maintenance risk of inconsistency.
Additional Locations (1)
- Add cleanup for requestAnimationFrame and setTimeout in scroll effect to prevent stale DOM writes on unmount or rapid selection changes - Fix disabled+selected card click: bounce the selection to re-trigger the engine auto-switch instead of deselecting Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
There was a problem hiding this comment.
🧹 Nitpick comments (1)
app/src/components/VoiceProfiles/ProfileCard.tsx (1)
44-49: Selection bounce pattern intentionally triggers engine auto-switch.The pattern of clearing
selectedProfileIdthen re-setting it viasetTimeout(..., 0)forces dependent effects (like engine auto-switch inFloatingGenerateBox.tsx) to re-run. This is a deliberate workaround for the "disabled but already selected" edge case.Note: This creates a brief intermediate state where
selectedProfileIdisnull, which will be observed by any component subscribed to the store. Ensure downstream components handle this gracefully.🔧 Alternative: Consider a dedicated "re-trigger" mechanism
Instead of bouncing through
null, you could add a counter or timestamp to the store that increments when a re-trigger is needed:+// In uiStore +profileSelectionVersion: 0, +bumpProfileSelectionVersion: () => set((s) => ({ profileSelectionVersion: s.profileSelectionVersion + 1 })), // In ProfileCard if (disabled && isSelected) { - setSelectedProfileId(null); - setTimeout(() => setSelectedProfileId(profile.id), 0); + bumpProfileSelectionVersion(); return; }Then effects can depend on both
selectedProfileIdandprofileSelectionVersionto re-run without the intermediate null state.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@app/src/components/VoiceProfiles/ProfileCard.tsx` around lines 44 - 49, The current "selection bounce" in ProfileCard (clearing selectedProfileId via setSelectedProfileId(null) then setSelectedProfileId(profile.id) in a setTimeout) causes a transient null state observed by subscribers; replace this with a dedicated re-trigger mechanism: add a profileSelectionVersion (or timestamp) in the same store as selectedProfileId, update it from ProfileCard instead of flipping to null (e.g., increment or set Date.now() when you need to re-trigger), and update dependent effects in FloatingGenerateBox (and any other subscribers reading selectedProfileId) to depend on both selectedProfileId and profileSelectionVersion so the engine auto-switch still re-runs without creating an intermediate null state.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@app/src/components/VoiceProfiles/ProfileCard.tsx`:
- Around line 44-49: The current "selection bounce" in ProfileCard (clearing
selectedProfileId via setSelectedProfileId(null) then
setSelectedProfileId(profile.id) in a setTimeout) causes a transient null state
observed by subscribers; replace this with a dedicated re-trigger mechanism: add
a profileSelectionVersion (or timestamp) in the same store as selectedProfileId,
update it from ProfileCard instead of flipping to null (e.g., increment or set
Date.now() when you need to re-trigger), and update dependent effects in
FloatingGenerateBox (and any other subscribers reading selectedProfileId) to
depend on both selectedProfileId and profileSelectionVersion so the engine
auto-switch still re-runs without creating an intermediate null state.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: fbac16c8-d799-4516-b606-400039fafc7b
📒 Files selected for processing (5)
app/src/components/Generation/FloatingGenerateBox.tsxapp/src/components/VoiceProfiles/ProfileCard.tsxapp/src/components/VoiceProfiles/ProfileList.tsxapp/src/lib/hooks/useGenerationForm.tsbackend/routes/profiles.py
1 participant
Summary
Test plan
🤖 Generated with Claude Code
Note
Medium Risk
Moderate risk due to changes in generation form defaulting and auto-switching engine based on selected profiles, which can affect what model/engine is used for requests. Backend impact appears minimal (route import/update timestamp behavior) but should be verified with preset voice listing and profile update flows.
Overview
Adds support for the new preset engine
qwen_custom_voiceand updates generation/profile selection so the form engine auto-aligns with the selected profile (including falling back away from preset engines when selecting non-preset voices).Updates the voice profile list UX to sort supported profiles first, gray out unsupported profiles instead of filtering them out, show an inline info hint when unsupported profiles exist, and auto-scroll the list to the currently selected profile after engine changes.
Initializes generation form
enginefrom the globalselectedEnginestore to prevent engine/profile desync when navigating between views, and includes a small backend tweak inprofiles.py(importsdatetime, used for updated timestamps in effects updates).Written by Cursor Bugbot for commit 7ebf57d. Configure here.
Summary by CodeRabbit
New Features
Bug Fixes