Added Parakeet as a STT model#766
Open
dustinwloring1988 wants to merge 1 commit into
Open
Conversation
Contributor
📝 WalkthroughWalkthroughAdds NVIDIA Parakeet TDT 0.6B (v2/v3) as a second STT engine alongside Whisper. A new canonical ChangesParakeet STT Engine Support
Dockerfile Frontend Build Refactor
Sequence Diagram(s)sequenceDiagram
participant Client
participant TranscriptionRoute as /transcribe endpoint
participant DB as capture_settings DB
participant STT_Registry as STT_HF_REPOS / normalize
participant TaskManager
participant STTBackend as get_stt_model() backend
Client->>TranscriptionRoute: POST /transcribe (file, language?, model?)
TranscriptionRoute->>DB: fetch saved stt_model (if model not provided)
DB-->>TranscriptionRoute: saved stt_model or default
TranscriptionRoute->>STT_Registry: normalize_stt_model_name(model)
STT_Registry-->>TranscriptionRoute: canonical model_name
TranscriptionRoute->>STTBackend: is_loaded() && model_name matches?
alt model not cached
TranscriptionRoute->>TaskManager: start background download task
TaskManager-->>Client: HTTP 202 (download in progress)
else model ready
TranscriptionRoute->>STTBackend: transcribe(audio_path, language, model_name)
STTBackend-->>TranscriptionRoute: transcription text
TranscriptionRoute-->>Client: 200 OK (text)
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 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}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 5
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
backend/mcp_server/tools.py (1)
324-324:⚠️ Potential issue | 🔴 Critical | ⚡ Quick winUndefined variable
model_sizewill cause NameError.The return statement references
model_sizebut the function definesmodel_name. This will crash at runtime.🐛 Proposed fix
return { "text": text, "duration": duration, "language": language, - "model": model_size, + "model": model_name, }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@backend/mcp_server/tools.py` at line 324, In the return statement at the specified location, the dictionary key "model" is assigned the undefined variable model_size, but the function actually defines model_name. Replace model_size with model_name in the "model" key assignment to fix the NameError that will occur at runtime.
🧹 Nitpick comments (2)
backend/services/captures.py (1)
121-122: 💤 Low valueConsider renaming the variable for clarity.
The variable
whispernow holds a generic STT backend that could be Parakeet. Consider renaming tosttorstt_backendfor clarity.♻️ Suggested rename
- whisper = get_stt_model() - resolved_stt = normalize_stt_model_name(stt_model or whisper.model_name) - transcript = await whisper.transcribe(str(audio_path), language, resolved_stt) + stt = get_stt_model() + resolved_stt = normalize_stt_model_name(stt_model or stt.model_name) + transcript = await stt.transcribe(str(audio_path), language, resolved_stt)Same change applies to
retranscribe_captureat lines 223-225.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@backend/services/captures.py` around lines 121 - 122, The variable name `whisper` is misleading because the result of `get_stt_model()` now represents a generic STT backend that could be Parakeet or other models, not specifically Whisper. Rename the variable `whisper` to `stt` or `stt_backend` throughout the code where it is assigned from `get_stt_model()` and update all its references including the usage in the `normalize_stt_model_name()` call. Apply this same variable rename change to the `retranscribe_capture` function at lines 223-225 where the identical pattern exists.app/src/components/CapturesTab/CapturesTab.tsx (1)
136-147: ⚡ Quick winTighten
formatSttModelNameinput type toSttModelId.Line 136 currently accepts
string, which bypasses compile-time guarantees for canonical STT ids and makes regressions easier to miss.Proposed refactor
+import type { SttModelId } from '`@/lib/api/types`'; ... -function formatSttModelName(modelName: string): string { +function formatSttModelName(modelName: SttModelId): string {🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@app/src/components/CapturesTab/CapturesTab.tsx` around lines 136 - 147, The formatSttModelName function currently accepts a generic string type for the modelName parameter, which lacks compile-time validation for canonical STT model IDs. Change the function signature to accept SttModelId type instead of string as the parameter type for modelName. This ensures that only valid STT model identifiers can be passed to the function and prevents potential regressions. The function implementation logic remains the same, only the input parameter type needs to be updated.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@backend/mcp_server/tools.py`:
- Around line 294-299: The database session in the block where `not model` is
checked is not properly closed because `next(get_db())` bypasses the generator's
cleanup mechanism. Replace the import of `get_db` with `SessionLocal` from the
database module, and wrap the database session creation and usage with a context
manager (with statement) to ensure the session is properly closed after calling
`settings_service.get_capture_settings(db)` to retrieve the stt_model.
In `@backend/models.py`:
- Around line 19-23: The decimal points in the Parakeet model IDs within the
STT_MODEL_PATTERN regex are unescaped, which allows invalid model names like
parakeet-tdt-0x6b-v2 to pass validation since unescaped dots match any
character. Escape the decimal points in both parakeet-tdt-0.6b-v2 and
parakeet-tdt-0.6b-v3 entries by prefixing each dot with a backslash to ensure
only literal dots are matched and prevent invalid model IDs from being persisted
through the CaptureSettingsUpdate model.
In `@backend/requirements.txt`:
- Around line 12-16: The transformers requirement on line 12
(transformers>=5.6.0) conflicts with the documented qwen-tts pin of
transformers==4.57.3, and since qwen-tts is actively used in the backend
(imported in pytorch_backend.py and qwen_custom_voice_backend.py), this version
mismatch must be resolved. Either verify the actual transformers version
constraint required by qwen-tts and update line 12 to be compatible with that
version, or update the qwen-tts installation method to use a version compatible
with transformers>=5.6.0, then update the comment on line 15 to accurately
reflect the pinned transformers version if environments are intentionally split.
In `@backend/routes/transcription.py`:
- Around line 44-49: The database session created with next(get_db()) is not
properly closed, causing connection leaks on every request where model is not
provided. Instead of using next(get_db()), refactor this block to either pass
the database session as a dependency parameter to this function or use a proper
context manager pattern with get_db() that ensures the generator's cleanup logic
(the finally block) executes to close the connection. Consider modifying the
function signature to accept a db parameter, or if that's not feasible, wrap the
get_db() call with a context manager that guarantees the session closure after
settings_service.get_capture_settings(db) completes and model is retrieved from
saved.stt_model.
In `@Dockerfile`:
- Around line 17-40: The Dockerfile uses heredoc syntax with the EOFPKG
delimiter to create a package.json file, but the Dockerfile parser cannot
interpret this without an explicit syntax directive. To fix this, add the syntax
directive `# syntax=docker/dockerfile:1.4+` as the first line of the Dockerfile
before any other instructions, which will enable support for heredoc syntax.
Alternatively, if you prefer to avoid adding the directive, refactor the RUN
command that creates the package-temp.json file to use printf or echo with
escaped newlines instead of heredoc syntax, which will work with the standard
Dockerfile parser without requiring the directive.
---
Outside diff comments:
In `@backend/mcp_server/tools.py`:
- Line 324: In the return statement at the specified location, the dictionary
key "model" is assigned the undefined variable model_size, but the function
actually defines model_name. Replace model_size with model_name in the "model"
key assignment to fix the NameError that will occur at runtime.
---
Nitpick comments:
In `@app/src/components/CapturesTab/CapturesTab.tsx`:
- Around line 136-147: The formatSttModelName function currently accepts a
generic string type for the modelName parameter, which lacks compile-time
validation for canonical STT model IDs. Change the function signature to accept
SttModelId type instead of string as the parameter type for modelName. This
ensures that only valid STT model identifiers can be passed to the function and
prevents potential regressions. The function implementation logic remains the
same, only the input parameter type needs to be updated.
In `@backend/services/captures.py`:
- Around line 121-122: The variable name `whisper` is misleading because the
result of `get_stt_model()` now represents a generic STT backend that could be
Parakeet or other models, not specifically Whisper. Rename the variable
`whisper` to `stt` or `stt_backend` throughout the code where it is assigned
from `get_stt_model()` and update all its references including the usage in the
`normalize_stt_model_name()` call. Apply this same variable rename change to the
`retranscribe_capture` function at lines 223-225 where the identical pattern
exists.
🪄 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: 7b41097d-8dff-4605-bd18-4431cce2be3c
📒 Files selected for processing (27)
Dockerfileapp/src/components/CapturesTab/CapturesTab.tsxapp/src/components/ServerSettings/ModelManagement.tsxapp/src/components/ServerTab/CapturesPage.tsxapp/src/components/VoiceProfiles/ProfileForm.tsxapp/src/components/VoiceProfiles/SampleUpload.tsxapp/src/i18n/locales/en/translation.jsonapp/src/i18n/locales/ja/translation.jsonapp/src/i18n/locales/zh-CN/translation.jsonapp/src/i18n/locales/zh-TW/translation.jsonapp/src/lib/api/client.tsapp/src/lib/api/types.tsapp/src/lib/hooks/useTranscription.tsbackend/backends/__init__.pybackend/backends/mlx_backend.pybackend/backends/pytorch_backend.pybackend/build_binary.pybackend/database/migrations.pybackend/database/models.pybackend/mcp_server/tools.pybackend/models.pybackend/requirements.txtbackend/routes/captures.pybackend/routes/transcription.pybackend/services/captures.pybackend/services/transcribe.pybackend/voicebox-server.spec
Comment on lines
+294
to
+299
| if not model: | ||
| from ..services import settings as settings_service | ||
| from ..database import get_db | ||
| db = next(get_db()) | ||
| saved = settings_service.get_capture_settings(db) | ||
| model = saved.stt_model if saved else None |
Contributor
There was a problem hiding this comment.
Database session is not properly closed.
Same issue as in transcription.py: using next(get_db()) bypasses the generator's cleanup. Consider using SessionLocal() directly with a context manager.
🐛 Proposed fix
if not model:
from ..services import settings as settings_service
- from ..database import get_db
- db = next(get_db())
- saved = settings_service.get_capture_settings(db)
- model = saved.stt_model if saved else None
+ from ..database import SessionLocal
+ with SessionLocal() as db:
+ saved = settings_service.get_capture_settings(db)
+ model = saved.stt_model if saved else None📝 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.
Suggested change
| if not model: | |
| from ..services import settings as settings_service | |
| from ..database import get_db | |
| db = next(get_db()) | |
| saved = settings_service.get_capture_settings(db) | |
| model = saved.stt_model if saved else None | |
| if not model: | |
| from ..services import settings as settings_service | |
| from ..database import SessionLocal | |
| with SessionLocal() as db: | |
| saved = settings_service.get_capture_settings(db) | |
| model = saved.stt_model if saved else None |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@backend/mcp_server/tools.py` around lines 294 - 299, The database session in
the block where `not model` is checked is not properly closed because
`next(get_db())` bypasses the generator's cleanup mechanism. Replace the import
of `get_db` with `SessionLocal` from the database module, and wrap the database
session creation and usage with a context manager (with statement) to ensure the
session is properly closed after calling
`settings_service.get_capture_settings(db)` to retrieve the stt_model.
Comment on lines
+19
to
+23
| STT_MODEL_PATTERN = ( | ||
| r"^(" | ||
| r"whisper-base|whisper-small|whisper-medium|whisper-large|whisper-turbo" | ||
| r"|parakeet-tdt-0.6b-v2|parakeet-tdt-0.6b-v3" | ||
| r"|base|small|medium|large|turbo" # legacy bare Whisper sizes |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
python - <<'PY'
import re
pattern = r"^(whisper-base|whisper-small|whisper-medium|whisper-large|whisper-turbo|parakeet-tdt-0.6b-v2|parakeet-tdt-0.6b-v3|base|small|medium|large|turbo)$"
tests = ["parakeet-tdt-0.6b-v2", "parakeet-tdt-0x6b-v2"]
for t in tests:
print(t, bool(re.fullmatch(pattern, t)))
PYRepository: jamiepine/voicebox
Length of output: 114
🏁 Script executed:
# Check the actual code in backend/models.py at the relevant lines
head -30 backend/models.py | tail -15Repository: jamiepine/voicebox
Length of output: 598
🏁 Script executed:
# Verify the pattern is indeed used at the mentioned line numbers
grep -n "STT_MODEL_PATTERN" backend/models.pyRepository: jamiepine/voicebox
Length of output: 392
🏁 Script executed:
# Check the context around Line 288 to confirm it's a settings update path
sed -n '280,295p' backend/models.pyRepository: jamiepine/voicebox
Length of output: 637
Escape the decimal points in Parakeet model IDs in the shared regex.
Line 22 uses 0.6b with unescaped dots, allowing invalid values like parakeet-tdt-0x6b-v2 to pass validation. These can be persisted via the CaptureSettingsUpdate model at line 288, causing failures during later registry lookups.
Proposed fix
STT_MODEL_PATTERN = (
r"^("
r"whisper-base|whisper-small|whisper-medium|whisper-large|whisper-turbo"
- r"|parakeet-tdt-0.6b-v2|parakeet-tdt-0.6b-v3"
+ r"|parakeet-tdt-0\.6b-v2|parakeet-tdt-0\.6b-v3"
r"|base|small|medium|large|turbo" # legacy bare Whisper sizes
r")$"
)📝 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.
Suggested change
| STT_MODEL_PATTERN = ( | |
| r"^(" | |
| r"whisper-base|whisper-small|whisper-medium|whisper-large|whisper-turbo" | |
| r"|parakeet-tdt-0.6b-v2|parakeet-tdt-0.6b-v3" | |
| r"|base|small|medium|large|turbo" # legacy bare Whisper sizes | |
| STT_MODEL_PATTERN = ( | |
| r"^(" | |
| r"whisper-base|whisper-small|whisper-medium|whisper-large|whisper-turbo" | |
| r"|parakeet-tdt-0\.6b-v2|parakeet-tdt-0\.6b-v3" | |
| r"|base|small|medium|large|turbo" # legacy bare Whisper sizes | |
| r")$" | |
| ) |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@backend/models.py` around lines 19 - 23, The decimal points in the Parakeet
model IDs within the STT_MODEL_PATTERN regex are unescaped, which allows invalid
model names like parakeet-tdt-0x6b-v2 to pass validation since unescaped dots
match any character. Escape the decimal points in both parakeet-tdt-0.6b-v2 and
parakeet-tdt-0.6b-v3 entries by prefixing each dot with a backslash to ensure
only literal dots are matched and prevent invalid model IDs from being persisted
through the CaptureSettingsUpdate model.
Comment on lines
+12
to
16
| transformers>=5.6.0 | ||
| accelerate>=0.26.0 | ||
| huggingface_hub>=0.20.0 | ||
| qwen-tts>=0.0.5 | ||
| # qwen-tts: installed separately with --no-deps (pins transformers==4.57.3) | ||
|
|
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== Dependency pins =="
rg -n --no-heading -C2 '(^|\s)(transformers|qwen-tts|qwen_tts)(==|>=|<=|~=|@|:| )|--no-deps' \
requirements.txt backend/requirements.txt Dockerfile README.md || true
echo
echo "== Runtime usage of qwen_tts =="
rg -n --no-heading -C2 '\bqwen_tts\b|qwen-tts' backend || true
echo
echo "== Build/bundle references =="
rg -n --no-heading -C2 'qwen_tts|qwen-tts|transformers' backend/build_binary.py backend/voicebox-server.spec || true
echo
echo "== Specifier overlap sanity check =="
python - <<'PY'
from packaging.specifiers import SpecifierSet
from packaging.version import Version
s_main = SpecifierSet(">=5.6.0")
s_qwen = SpecifierSet("==4.57.3")
v = Version("4.57.3")
print("4.57.3 satisfies main spec:", v in s_main)
print("4.57.3 satisfies qwen spec:", v in s_qwen)
PYRepository: jamiepine/voicebox
Length of output: 18323
Resolve the transformers version conflict between direct dependency and qwen-tts pin.
Line 12 specifies transformers>=5.6.0, but Line 15 documents qwen-tts as pinning transformers==4.57.3. Version 4.57.3 does not satisfy the >=5.6.0 requirement, and since qwen_tts is actively used in the backend (imported in pytorch_backend.py and qwen_custom_voice_backend.py) and bundled in the binary build, this creates a version conflict. Clarify whether environments are intentionally split, document the required version constraints, or adjust the floor to be compatible.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@backend/requirements.txt` around lines 12 - 16, The transformers requirement
on line 12 (transformers>=5.6.0) conflicts with the documented qwen-tts pin of
transformers==4.57.3, and since qwen-tts is actively used in the backend
(imported in pytorch_backend.py and qwen_custom_voice_backend.py), this version
mismatch must be resolved. Either verify the actual transformers version
constraint required by qwen-tts and update line 12 to be compatible with that
version, or update the qwen-tts installation method to use a version compatible
with transformers>=5.6.0, then update the comment on line 15 to accurately
reflect the pinned transformers version if environments are intentionally split.
Comment on lines
+44
to
+49
| if not model: | ||
| from ..services import settings as settings_service | ||
| from ..database import get_db | ||
| db = next(get_db()) | ||
| saved = settings_service.get_capture_settings(db) | ||
| model = saved.stt_model if saved else None |
Contributor
There was a problem hiding this comment.
Database session is not properly closed.
Using next(get_db()) bypasses the generator's finally block that closes the session. This leaks database connections on every request where model is not provided.
🐛 Proposed fix using context manager pattern
# Resolve model: explicit param > user's saved setting > backend default
if not model:
from ..services import settings as settings_service
from ..database import get_db
- db = next(get_db())
- saved = settings_service.get_capture_settings(db)
- model = saved.stt_model if saved else None
+ db_gen = get_db()
+ db = next(db_gen)
+ try:
+ saved = settings_service.get_capture_settings(db)
+ model = saved.stt_model if saved else None
+ finally:
+ try:
+ next(db_gen)
+ except StopIteration:
+ passOr more cleanly, use contextlib:
from contextlib import contextmanager
# In the function:
if not model:
from ..services import settings as settings_service
from ..database import get_db
from contextlib import closing
with closing(next(iter_db := get_db())) as db:
# Note: This still won't work correctly. Better approach:
passThe cleanest fix is to extract a helper or pass db as a dependency:
+from ..database import SessionLocal
+
# In the function:
if not model:
from ..services import settings as settings_service
- from ..database import get_db
- db = next(get_db())
- saved = settings_service.get_capture_settings(db)
- model = saved.stt_model if saved else None
+ with SessionLocal() as db:
+ saved = settings_service.get_capture_settings(db)
+ model = saved.stt_model if saved else None🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@backend/routes/transcription.py` around lines 44 - 49, The database session
created with next(get_db()) is not properly closed, causing connection leaks on
every request where model is not provided. Instead of using next(get_db()),
refactor this block to either pass the database session as a dependency
parameter to this function or use a proper context manager pattern with get_db()
that ensures the generator's cleanup logic (the finally block) executes to close
the connection. Consider modifying the function signature to accept a db
parameter, or if that's not feasible, wrap the get_db() call with a context
manager that guarantees the session closure after
settings_service.get_capture_settings(db) completes and model is retrieved from
saved.stt_model.
Comment on lines
+17
to
+40
| RUN cat > package-temp.json << 'EOFPKG' && mv package-temp.json package.json | ||
| { | ||
| "name": "voicebox-container", | ||
| "version": "0.5.0", | ||
| "private": true, | ||
| "workspaces": [ | ||
| "app", | ||
| "web" | ||
| ], | ||
| "scripts": { | ||
| "build:web": "cd web && bun run build" | ||
| }, | ||
| "dependencies": { | ||
| "loaders.css": "^0.1.2", | ||
| "react-loaders": "^3.0.1" | ||
| }, | ||
| "devDependencies": { | ||
| "@biomejs/biome": "2.3.12", | ||
| "@types/node": "^20.0.0", | ||
| "tailwindcss": "^4.1.18", | ||
| "typescript": "^5.6.0" | ||
| } | ||
| } | ||
| EOFPKG |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Verify parser compatibility signals for this Dockerfile change.
set -euo pipefail
echo "== Dockerfile first 5 lines (check syntax directive) =="
sed -n '1,5p' Dockerfile
echo
echo "== Locate the heredoc RUN block =="
rg -n -C3 "package-temp.json|EOFPKG|RUN cat >" Dockerfile || true
echo
echo "== Hadolint parse check (if installed) =="
if command -v hadolint >/dev/null 2>&1; then
hadolint Dockerfile || true
else
echo "hadolint not installed in this environment."
fiRepository: jamiepine/voicebox
Length of output: 1045
Heredoc JSON block requires explicit syntax directive or parser-safe rewrite.
Hadolint reports a DL1000 parse error on line 18 (unexpected '{'). The Dockerfile uses heredoc syntax (<< 'EOFPKG') without the required # syntax=docker/dockerfile:1.4+ directive at the file top, causing standard parsers to misinterpret the JSON block as a Docker instruction. This blocks the build at the parse phase.
Add the syntax directive at the start of the file, or refactor to avoid heredoc syntax:
Option 1: Add syntax directive (1 line)
+# syntax=docker/dockerfile:1.4+
# ============================================================
# Voicebox — Local TTS Server with Web UI (CPU)Option 2: Refactor with printf (parser-safe, no directive needed)
-RUN cat > package-temp.json << 'EOFPKG' && mv package-temp.json package.json
-{
- "name": "voicebox-container",
- "version": "0.5.0",
- "private": true,
- "workspaces": [
- "app",
- "web"
- ],
- "scripts": {
- "build:web": "cd web && bun run build"
- },
- "dependencies": {
- "loaders.css": "^0.1.2",
- "react-loaders": "^3.0.1"
- },
- "devDependencies": {
- "`@biomejs/biome`": "2.3.12",
- "`@types/node`": "^20.0.0",
- "tailwindcss": "^4.1.18",
- "typescript": "^5.6.0"
- }
-}
-EOFPKG
+RUN printf '%s\n' \
+'{' \
+' "name": "voicebox-container",' \
+' "version": "0.5.0",' \
+' "private": true,' \
+' "workspaces": ["app", "web"],' \
+' "scripts": { "build:web": "cd web && bun run build" },' \
+' "dependencies": {' \
+' "loaders.css": "^0.1.2",' \
+' "react-loaders": "^3.0.1"' \
+' },' \
+' "devDependencies": {' \
+' "`@biomejs/biome`": "2.3.12",' \
+' "`@types/node`": "^20.0.0",' \
+' "tailwindcss": "^4.1.18",' \
+' "typescript": "^5.6.0"' \
+' }' \
+'}' > package.json📝 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.
Suggested change
| RUN cat > package-temp.json << 'EOFPKG' && mv package-temp.json package.json | |
| { | |
| "name": "voicebox-container", | |
| "version": "0.5.0", | |
| "private": true, | |
| "workspaces": [ | |
| "app", | |
| "web" | |
| ], | |
| "scripts": { | |
| "build:web": "cd web && bun run build" | |
| }, | |
| "dependencies": { | |
| "loaders.css": "^0.1.2", | |
| "react-loaders": "^3.0.1" | |
| }, | |
| "devDependencies": { | |
| "@biomejs/biome": "2.3.12", | |
| "@types/node": "^20.0.0", | |
| "tailwindcss": "^4.1.18", | |
| "typescript": "^5.6.0" | |
| } | |
| } | |
| EOFPKG | |
| # syntax=docker/dockerfile:1.4+ | |
| # ============================================================ | |
| # Voicebox — Local TTS Server with Web UI (CPU) |
Suggested change
| RUN cat > package-temp.json << 'EOFPKG' && mv package-temp.json package.json | |
| { | |
| "name": "voicebox-container", | |
| "version": "0.5.0", | |
| "private": true, | |
| "workspaces": [ | |
| "app", | |
| "web" | |
| ], | |
| "scripts": { | |
| "build:web": "cd web && bun run build" | |
| }, | |
| "dependencies": { | |
| "loaders.css": "^0.1.2", | |
| "react-loaders": "^3.0.1" | |
| }, | |
| "devDependencies": { | |
| "@biomejs/biome": "2.3.12", | |
| "@types/node": "^20.0.0", | |
| "tailwindcss": "^4.1.18", | |
| "typescript": "^5.6.0" | |
| } | |
| } | |
| EOFPKG | |
| RUN printf '%s\n' \ | |
| '{' \ | |
| ' "name": "voicebox-container",' \ | |
| ' "version": "0.5.0",' \ | |
| ' "private": true,' \ | |
| ' "workspaces": ["app", "web"],' \ | |
| ' "scripts": { "build:web": "cd web && bun run build" },' \ | |
| ' "dependencies": {' \ | |
| ' "loaders.css": "^0.1.2",' \ | |
| ' "react-loaders": "^3.0.1"' \ | |
| ' },' \ | |
| ' "devDependencies": {' \ | |
| ' "`@biomejs/biome`": "2.3.12",' \ | |
| ' "`@types/node`": "^20.0.0",' \ | |
| ' "tailwindcss": "^4.1.18",' \ | |
| ' "typescript": "^5.6.0"' \ | |
| ' }' \ | |
| '}' > package.json |
🧰 Tools
🪛 Hadolint (2.14.0)
[error] 18-18: unexpected '{'
expecting '#', '', ADD, ARG, CMD, COPY, ENTRYPOINT, ENV, EXPOSE, FROM, HEALTHCHECK, LABEL, MAINTAINER, ONBUILD, RUN, SHELL, STOPSIGNAL, USER, VOLUME, WORKDIR, a pragma, at least one space, or end of input
(DL1000)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@Dockerfile` around lines 17 - 40, The Dockerfile uses heredoc syntax with the
EOFPKG delimiter to create a package.json file, but the Dockerfile parser cannot
interpret this without an explicit syntax directive. To fix this, add the syntax
directive `# syntax=docker/dockerfile:1.4+` as the first line of the Dockerfile
before any other instructions, which will enable support for heredoc syntax.
Alternatively, if you prefer to avoid adding the directive, refactor the RUN
command that creates the package-temp.json file to use printf or echo with
escaped newlines instead of heredoc syntax, which will work with the standard
Dockerfile parser without requiring the directive.
Source: Linters/SAST tools
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds support for NVIDIA Parakeet speech-to-text integration in Voicebox.
What's Included
Notes
Summary by CodeRabbit
Release Notes
New Features
whisper-turbo,parakeet-tdt-0.6b-v2)Bug Fixes
turbotowhisper-turboLocalization