Added features: local transcription, multi-language support, preview, customizable styles by ollisulopuisto · Pull Request #7 · riseandignite/capslap

ollisulopuisto · 2026-01-30T13:47:46Z

Added features: local transcription, multi-language support, preview, customizable styles, editing and saving subs

…ting and HDR video support, refactor concurrent task management."

…nd caption burning RPCs with a new CaptionEditor UI.

…nce Whisper word parsing by merging sub-word tokens.

…g, and enhance caption word wrapping with hyphen splitting and improved layout.

…e segment/word time range detection.

…ing new Rust types and RPC, and remove the save icon from the UI.

… the burn command.

…d on length and add a corresponding test case.

…ific sidecar files.

… and add double-click to edit captions.

…w Rust RPC method and Electron UI.

…in preview, and display playback error messages for unsupported formats.

…Editor.

… bump version to 1.0.5

… to 1.0.6

Copilot

Pull request overview

This pull request adds significant new functionality to CapSlap, a video caption generation tool. The changes introduce local transcription via whisper.cpp (eliminating cloud API dependency), a caption editor with live preview, multi-language support, expanded font library (30+ fonts), and various UI/UX improvements.

Changes:

Local transcription using whisper.cpp with caching and multiple model support (including new "turbo" model)
Caption editor component with word-level editing, segment shifting, and live preview generation
Expanded font library organized into 5 categories with 30+ fonts, plus font download automation
Enhanced CI/CD with release workflow and platform-specific binary management
UI improvements including font size control, output resolution selection, crop strategies, and progress indicators

Reviewed changes

Copilot reviewed 29 out of 64 changed files in this pull request and generated 15 comments.

Show a summary per file

File	Description
scripts/download_fonts.sh	New script to download Google Fonts - has syntax error
scripts/download-binaries.sh	CI script for downloading platform-specific FFmpeg and whisper binaries - has architecture bug
rust/src/whisper.rs	Core transcription logic with local whisper.cpp support, model fallback chain, caching, and word merging - has duplicate comments and unused variable
rust/src/types.rs	Extended type definitions for new features (preview, burn, save/load captions)
rust/src/rpc.rs	Added comprehensive unit tests for RPC layer
rust/src/audio.rs	Minor formatting improvements
rust/src/bin/debug_ffmpeg.rs	New debugging utility for FFmpeg detection
rust/src/bin/core.rs	Enhanced RPC handler with cancellation support and new methods (transcribe, burn, preview, save/load)
electron/package.json	Version bump to 1.0.6, added dependencies for testing and UI components
electron/vitest.config.ts	New test configuration with coverage setup
electron/test/setup.ts	Test environment setup with mocked Electron APIs
electron/lib/utils.test.ts	Comprehensive tests for className utility
electron/lib/caption-utils.ts	New utility functions for caption formatting and manipulation
electron/lib/caption-utils.test.ts	Tests for caption utilities
electron/lib/preload/*	Updated API to support request IDs for cancellation
electron/lib/main/sidecar.ts	Enhanced logging and cancellation support
electron/lib/main/app.ts	Extended resource protocol to support local files and more video formats
electron/eslint.config.mjs	New ESLint configuration
electron/electron-builder.yml	Added Linux binary support
electron/app/components/ui/*	New UI components (Textarea, Dropdown, updated Slider/Separator)
electron/app/components/ModelDownloader.tsx	Added "turbo" model support
electron/app/components/CaptionEditor.tsx	New caption editor with segment editing and word shifting - has unused variables
electron/app/app.tsx	Major UI overhaul with editor integration, preview generation, and expanded settings - has duplicate field assignments
.github/workflows/*	New CI and release workflows for automated builds
README.md	Updated documentation highlighting offline capability and new features
.gitignore	Expanded to exclude build artifacts and binaries

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-30T13:53:17Z

scripts/download_fonts.sh

+# Cinzel
+download_google_font "Cinzel" "Bold"
+# Bodoni Moda
+download_google_font "Bodoni Moda" "Bold" // Note: Variable font usually


Invalid comment syntax in shell script. Shell scripts use '#' for comments, not '//'. This will cause a syntax error when the script is executed.

Copilot · 2026-01-30T13:53:17Z

rust/src/whisper.rs

+/// Get possible project FFmpeg binary paths
 /// Get possible project FFmpeg binary paths


Duplicate docstring comment. Line 533 and 534 both have the same comment "/// Get possible project FFmpeg binary paths". Remove the duplicate on line 533.

Copilot · 2026-01-30T13:53:17Z

rust/src/whisper.rs

+/// Get possible project ffprobe binary paths
 /// Get possible project ffprobe binary paths


Duplicate docstring comment. Line 596 and 597 both have the same comment "/// Get possible project ffprobe binary paths". Remove the duplicate on line 596.

Copilot · 2026-01-30T13:53:18Z

electron/app/app.tsx

+          fontName: getFontName(videoSettings.selectedFont),
+          textColor: videoSettings.textColor,
+          highlightWordColor: videoSettings.highlightWordColor,
+          outlineColor: videoSettings.outlineColor,
+          glowEffect: videoSettings.glowEffect,
+          position: videoSettings.captionPosition,
+          outputSize: videoSettings.outputSize,
+          position: videoSettings.captionPosition,
+          outputSize: videoSettings.outputSize,


Duplicate field assignment. The position field is assigned twice on lines 842-843. Remove the duplicate assignment on line 842.

Suggested change

fontName: getFontName(videoSettings.selectedFont),

textColor: videoSettings.textColor,

highlightWordColor: videoSettings.highlightWordColor,

outlineColor: videoSettings.outlineColor,

glowEffect: videoSettings.glowEffect,

position: videoSettings.captionPosition,

outputSize: videoSettings.outputSize,

position: videoSettings.captionPosition,

outputSize: videoSettings.outputSize,

fontName: getFontName(videoSettings.selectedFont,

textColor: videoSettings.textColor,

highlightWordColor: videoSettings.highlightWordColor,

outlineColor: videoSettings.outlineColor,

glowEffect: videoSettings.glowEffect,

position: videoSettings.captionPosition,

outputSize: videoSettings.outputSize,

Copilot · 2026-01-30T13:53:18Z

electron/app/app.tsx

+          glowEffect: videoSettings.glowEffect,
+          position: videoSettings.captionPosition,
+          outputSize: videoSettings.outputSize,
+          outputSize: videoSettings.outputSize,


Duplicate field assignment. The outputSize field is assigned twice on lines 929-930. Remove the duplicate assignment on line 929.

Suggested change

outputSize: videoSettings.outputSize,

Copilot · 2026-01-30T13:53:19Z

scripts/download-binaries.sh

+    ARCH=$(uname -m)
+    if [[ "$ARCH" == "arm64" ]]; then
+        WHISPER_ASSET="whisper-cli-macos-arm64"
+    else
+        WHISPER_ASSET="whisper-cli-macos-x64"
+    fi
+
+    curl -L "https://github.com/ggerganov/whisper.cpp/releases/download/${WHISPER_VERSION}/${WHISPER_ASSET}" \
+        -o "$BIN_DIR/whisper-cli-macos-arm64"
+    chmod +x "$BIN_DIR/whisper-cli-macos-arm64"


Incorrect Linux binary paths for macOS build. In the download_macos function (lines 16-52), the whisper binary is always saved as "whisper-cli-macos-arm64" regardless of the architecture detected. On line 48, both arm64 and x64 architectures result in the same output filename. The x64 binary should be saved as "whisper-cli-macos-x64" when ARCH is x64.

Copilot · 2026-01-30T13:53:20Z

electron/app/components/CaptionEditor.tsx

+const getFontName = (fontId: string): string => {
+  return FONT_NAMES[fontId as keyof typeof FONT_NAMES] || 'Montserrat Black'
+}
+


Unused variable getFontName.

Suggested change

const getFontName = (fontId: string): string => {

return FONT_NAMES[fontId as keyof typeof FONT_NAMES] || 'Montserrat Black'

}

Copilot · 2026-01-30T13:53:20Z

electron/app/app.tsx

+  SelectGroup,
+  SelectLabel,


Unused imports SelectGroup, SelectLabel.

Suggested change

SelectGroup,

SelectLabel,

Copilot · 2026-01-30T13:53:20Z

electron/app/app.tsx

+  DropdownMenuLabel,
+  DropdownMenuSeparator,
+  DropdownMenuTrigger,
+  DropdownMenuSub,
+  DropdownMenuSubTrigger,
+  DropdownMenuSubContent,
+  DropdownMenuGroup,


Unused imports DropdownMenuGroup, DropdownMenuLabel, DropdownMenuSeparator.

Suggested change

DropdownMenuLabel,

DropdownMenuSeparator,

DropdownMenuTrigger,

DropdownMenuSub,

DropdownMenuSubTrigger,

DropdownMenuSubContent,

DropdownMenuGroup,

DropdownMenuTrigger,

DropdownMenuSub,

DropdownMenuSubTrigger,

DropdownMenuSubContent,

Copilot · 2026-01-30T13:53:20Z

scripts/download-binaries.sh

+    # Download FFmpeg from gyan.dev (Windows builds with all codecs)
+    echo "Downloading FFmpeg..."
+    FFMPEG_URL="https://github.com/BtbN/FFmpeg-Builds/releases/download/latest/ffmpeg-master-latest-win64-gpl.zip"
+    curl -L "$FFMPEG_URL" -o /tmp/ffmpeg-win.zip
+    unzip -o /tmp/ffmpeg-win.zip -d /tmp/ffmpeg-win
+
+    # Find and copy binaries
+    find /tmp/ffmpeg-win -name "ffmpeg.exe" -exec cp {} "$BIN_DIR/" \;
+    find /tmp/ffmpeg-win -name "ffprobe.exe" -exec cp {} "$BIN_DIR/" \;
+
+    # Download whisper.cpp Windows build
+    echo "Downloading whisper-cli..."
+    curl -L "https://github.com/ggerganov/whisper.cpp/releases/download/${WHISPER_VERSION}/whisper-cli-win64.exe" \
+        -o "$BIN_DIR/whisper-cli.exe"


This script downloads and executes precompiled ffmpeg and whisper-cli binaries from third-party GitHub URLs (including the mutable latest FFmpeg builds) without any checksum, signature, or pinning to an immutable identifier. If the upstream project, tag, or delivery path is compromised (or a MITM/TLS termination point is abused), a malicious binary could be fetched and then executed in CI or shipped to users, leading to remote code execution. To mitigate this, pin these downloads to immutable artifacts and verify their integrity (e.g., using published checksums/signatures or vendoring known-good binaries) instead of relying on latest and unverified downloads.

- scripts/download_fonts.sh: Fix invalid shell comment syntax (// -> #) - scripts/download-binaries.sh: Fix architecture bug - save whisper binary with correct filename for both arm64 and x64 - rust/src/whisper.rs: Remove duplicate docstrings for get_project_ffmpeg_paths and get_project_ffprobe_paths - electron/app/components/CaptionEditor.tsx: Remove unused FONT_NAMES and getFontName - electron/app/app.tsx: Remove unused imports (SelectGroup, SelectLabel, DropdownMenuLabel, DropdownMenuSeparator, DropdownMenuGroup) and fix duplicate field assignments

…form resources

ollisulopuisto added 30 commits December 9, 2025 22:20

chore: update gitignore and project configuration

39ca1db

feat: Introduce Storyteller caption style with multi-line text format…

07fbe4b

…ting and HDR video support, refactor concurrent task management."

feat: Add task cancellation, video frame extraction, transcription, a…

705d4d7

…nd caption burning RPCs with a new CaptionEditor UI.

feat: Refine caption line wrapping minimum width calculation and enha…

611406b

…nce Whisper word parsing by merging sub-word tokens.

feat: Add Textarea component, support local and MOV video file loadin…

c570786

…g, and enhance caption word wrapping with hyphen splitting and improved layout.

fix: Remove redundant percentage from export status display.

cb12746

feat: add video poster support to the caption editor and refine activ…

698b940

…e segment/word time range detection.

vertical pos adjust

c5937fa

feat: Add backend-driven caption layout preview to the editor, utiliz…

07647df

…ing new Rust types and RPC, and remove the save icon from the UI.

fix: Round segment and word timestamps to the nearest millisecond for…

44c6fd4

… the burn command.

feat: proportionally distribute time spans for hyphenated tokens base…

b96b7bb

…d on length and add a corresponding test case.

feat: Implement saving and loading of adjusted captions to video-spec…

a367e71

…ific sidecar files.

feat: Round segment and word timestamps to integers for backend calls…

19eed94

… and add double-click to edit captions.

feat: Implement dynamic caption previews for templates by adding a ne…

4057f30

…w Rust RPC method and Electron UI.

feat: Enhance video format support, reflect current caption settings …

5137d51

…in preview, and display playback error messages for unsupported formats.

refactor: remove video playback and live caption preview from Caption…

7d780e6

…Editor.

FFmpeg missing ASS

6f191f3

feat: Add FFmpeg libass availability check and a subtitle test asset.

9c54d53

buncha font work

be71f07

Add CI/CD release pipeline

9d4742d

readme updates

76fe04b

buncha cleanup pre-testing

ceea061

testing

2158b6c

fix failing test

c2e7f57

chore: update bun.lock and fix lint setup to resolve CI error

ba3aade

CI/CD fixes

41eff02

chore: bump version to 1.0.3

b083370

chore: update release workflow and bump version to 1.0.4

1081cce

fix(ci): scope linux ffmpeg extraction to avoid permission errors and…

c9b5f54

… bump version to 1.0.5

fix(build): disable code signing for macos ci builds and bump version…

0fefb53

… to 1.0.6

Copilot AI review requested due to automatic review settings January 30, 2026 13:47

Copilot started reviewing on behalf of ollisulopuisto January 30, 2026 13:48 View session

Copilot AI reviewed Jan 30, 2026

View reviewed changes

ollisulopuisto added 10 commits January 30, 2026 16:51

del old file

437e060

chore: bump version to 1.0.7

8236a19

ci: add electron build check to CI workflow

a7c736b

chore: bump version to 1.0.8

5a18016

ci: fix mac build by adding arch to artifact name and separating plat…

bf3cf2e

…form resources

feat: add manual preview refresh button to CaptionEditor

1ff7aa7

chore: exclude large models from bundle to reduce binary size

60cd91e

chore: bump version to 1.0.9

9218042

fix: use reliable FFmpeg source for Linux builds

ffb53e1

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Added features: local transcription, multi-language support, preview, customizable styles#7

Added features: local transcription, multi-language support, preview, customizable styles#7
ollisulopuisto wants to merge 40 commits intoriseandignite:mainfrom
ollisulopuisto:main

ollisulopuisto commented Jan 30, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Copilot AI Jan 30, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		/// Get possible project FFmpeg binary paths
		/// Get possible project FFmpeg binary paths

		/// Get possible project ffprobe binary paths
		/// Get possible project ffprobe binary paths

	const getFontName = (fontId: string): string => {
	return FONT_NAMES[fontId as keyof typeof FONT_NAMES] \|\| 'Montserrat Black'
	}

Conversation

ollisulopuisto commented Jan 30, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants