FuseClip 0.75.1

FuseClip v0.75.1 is a maintenance release: it fixes Manus sign-in and chat on Windows ARM64, repairs the in-app update flow, and refreshes the app icons and Windows installer.

Improvements

• Refreshed icons: New app and installer icons, with a redesigned Windows installer wizard.
• Single progress indicator: Installs and updates now show one overall progress percentage instead of several.

Fixes

• Fixed Manus failing to work on Windows ARM64.
• Fixed Manus chat and sign-in handling issues.
• Fixed problems that could occur while updating the app.
• Fixed broken ffmpeg source for Windows ARM64.
• Fixed an audio (Opus) issue.

Install

Download the installer for your platform from the FuseClip releases page.

• macOS (Apple Silicon): Open the .dmg and drag FuseClip into Applications.
• Windows: Run the x64 installer for Intel/AMD PCs or the arm64 installer for ARM PCs.

Already running FuseClip? The app will offer an update once v0.75.1 is published.

FuseClip 0.75.0

FuseClip v0.75.0 rebuilds face tracking so the crop follows the right person through every camera cut, switches layouts only when it should, and moves like a camera operator instead of jittering.

New

• Speaker-aware framing: When two people share one shot but are too close for split screen, the crop now cuts to whoever is talking, using the transcript's speaker labels matched to whose mouth is moving.
• Frame-accurate cut detection: Camera cuts are now found on the exact source frame, so layouts and crops change at the cut instead of a fraction of a second before or after it.

Improvements

• Person tracking across shots: Each person is followed through the whole shot, so a missed detection or a brief turn of the head no longer throws the crop to someone else.
• Steadier layouts: Split-screen and multi-person layouts only change when someone joins or leaves for good. A person walking through the frame or a momentary missed detection no longer causes a quick switch and back.
• Smoother camera movement: The crop holds still while the speaker stays roughly in place, and glides to a new position only when they settle somewhere else.
• Logo and badge filtering: Channel logos, avatar badges and other graphics fixed in a corner are no longer mistaken for a second person.
• Interface polish: Tidied spacing, header heights and button sizing across the editing panels.

Fixes

• Fixed face tracking falling back to a static center crop for an entire clip when it contained back-to-back crowded shots, such as fast-cut montage intros.

Install

Download the installer for your platform from the FuseClip releases page.

• macOS (Apple Silicon): Open the .dmg and drag FuseClip into Applications.
• Windows: Run the x64 installer for Intel/AMD PCs or the arm64 installer for ARM PCs.

Already running FuseClip? The app will offer an update once v0.75.0 is published.

FuseClip 0.74.1

FuseClip v0.74.1 improves Manus login, clip editing performance, caption previews, and face-tracking stability.

Improvements

• Manus onboarding: Complete Manus onboarding and phone-number verification directly within FuseClip’s login window.
• Clip editing performance: Improved responsiveness when adding clips to the timeline.

Fixes

• Manus button detection: Fixed detection of the Send button so AI requests submit reliably.
• Caption preset previews: Fixed cropped text and improved spacing, font compatibility, and emphasized-word previews across languages.
• Next button: Fixed advancing from caption editing to existing results without unnecessarily restarting processing.
• Caption editor: Fixed caption scrolling when switching between clips.
• Face tracking: Improved framing for different source aspect ratios, reduced false face detections, and stabilized layouts when a face is briefly obscured or turned away.

Install

Download the installer for your platform from the [FuseClip releases page](https://github.com/Amfaiz/fuseclip/releases).

• macOS (Apple Silicon): Open the .dmg and drag FuseClip into Applications.
• Windows: Run the x64 installer for Intel/AMD PCs or the arm64 installer for ARM PCs.

Already running FuseClip? The app will offer an update once v0.74.1 is published.

FuseClip 0.74.0

FuseClip v0.74.0 brings a more flexible clip editor, new viral caption styles, music and overlays, tracking for up to four faces, and more control over AI clipping—with a refreshed interface and smoother video playback.

New

• Clip timeline: Extend or shorten generated clips, adjust their position, manually add your own clips, and remove unwanted ones on a zoomable timeline.
• Custom clip length and count: Define your desired clip length and number of clips when generating clips with AI.
• Multi-face tracking: Track up to four faces and arrange multiple speakers in split-screen layouts.
• Word-by-word caption editor: Edit individual words, delete words or caption rows, move words between rows, and choose which words to emphasize directly within the clip editing step.
• Viral caption presets: Added glowing captions, emphasized-word styles, and more presets, including Headline, Orbit, Bayan, Uchhaal, and Lafz.
• Caption layout controls: Add custom breaks and set the number of words per line, maximum characters per line, and number of lines to shape captions for your video.
• AI-generated titles and hashtags: Automatically generate titles and hashtags for each clip to help prepare it for publishing.
• Background music: Add background music to your clips.
• Logo and watermark: Add a logo overlay or watermark to brand your videos.
• Video overlays: Add an overlay over your video to customize its look.
• Manus integration: Connect your Manus account to power AI clip detection, caption processing, and clip titles and hashtags. Reconnect from Settings when your session expires.
• OpenAI-compatible APIs: Use your own Chat Completions endpoint with a custom base URL, model, and optional API key. Test the connection directly from AI Provider settings.
• Custom clipping prompts: Edit the system prompts used for viral clip discovery and quality review.

Improvements

• Streaming video playback: Videos now stream for playback instead of loading the entire file into memory.
• Refreshed workspace: Updated the sidebar and editing workflow to make project navigation, clip review, and caption preparation easier to follow.
• Windows GPU transcription: Added a bundled Vulkan transcription engine for supported AMD and Intel GPUs, with CPU fallback in Auto mode when acceleration fails.
• Selective AI processing: Prepare captions and publishing copy for new or resized clips, with controls to rerun caption processing stages when needed.
• Caption customization: Added more font controls and refined multilingual caption layout, timing, and rendering.
• Better error handling: Clearer error messages and recovery guidance for failed requests, authentication problems, and provider errors, plus retries for temporary network failures and rate limits.

Fixes

• Fixed AI requests continuing to wait when Manus credits are exhausted.
• Fixed flicker when switching views during face tracking and split-screen playback.
• Fixed blank video previews caused by incorrect video filename parsing.
• Reduced preview flicker during playback and clip editing.
• Fixed saving clips at the source video's original aspect ratio.
• Improved export encoder detection and fallback handling when an H.264 encoder is unavailable.
• Fixed the sign-in window showing the default Electron icon.

Install

Download the installer for your platform from the FuseClip releases page.

• macOS (Apple Silicon): Open the .dmg and drag FuseClip into Applications.
• Windows: Run the x64 installer for Intel/AMD PCs or the arm64 installer for ARM PCs.

Already running FuseClip? The app will offer an update once v0.74.0 is published.

FuseClip 0.73.0

FuseClip v0.73.0 introduces custom font management, a full linear and radial gradient styling engine, multilingual AI script normalization for Hindi, Urdu, and Arabic, intelligent Gemini rate limiting, and major Windows video encoding and installer improvements.

New

• Custom Font Management: Install .ttf, .otf, .woff, and .woff2 fonts directly through a new Fonts tab in Settings. Custom fonts load dynamically into the renderer and are selectable across all caption presets.
• Linear & Radial Gradient Engine: Style captions with vibrant gradients using an interactive color picker featuring angle presets (0°–315°), custom color stops, and opacity blending across text, highlights, and plates.
• Multilingual Script Normalization: Added a Native vs. Roman script selector for Hindi and Urdu, paired with Gemini AI post-processing that seamlessly converts English loanwords into natural native script while preserving microsecond word timings.
• Gemini Rate Limiting & Quota Resilience: Built-in request pacing (20 RPM / 200k TPM ceilings), exponential backoff, and Google RPC retryDelay handling to safely prevent 429 rate limit errors during AI clipping and script fixing.

Improvements & Fixes

• Windows 9:16 Export Fix: Implemented pre-flight capability probing and automatic software fallbacks to resolve Windows Media Foundation (MFT) crashes on vertical/portrait (1080x1920) video exports.
• Precision Caption Timing: Added an 80ms audio onset lead so captions appear naturally as speech begins, eliminated word transition dead zones, and smoothed cue endings.
• Zero-UAC Windows Installer: Migrated to a per-user installation (%LocalAppData%) for seamless auto-updates without Windows Admin UAC prompts, including automatic cleanup and migration of legacy Program Files installs.
• Filesystem Fallbacks: Guarded against crashes when Windows Known Folders (FOLDERID_Videos) are unlinked or missing, safely falling back to user home or document directories.

FuseClip 0.72.0

FuseClip v0.72.0

FuseClip v0.72.0 introduces the AI-Powered Viral Clipper powered by Google Gemini, an interactive in-app Captions & Transcript Editor with live re-burning, an overhauled Face Tracking and 9:16 Split-Screen engine with ONNX detection, automatic project thumbnail generation with disk synchronization, and major Windows performance optimizations.

What's New

AI-Powered Viral Clipper (Google Gemini)

• Automated Short-Form Viral Clipping: Automatically detects engaging, high-CTR moments from long-form video transcripts and generates ready-to-share clips optimized for TikTok, Instagram Reels, and YouTube Shorts.
• Multi-Factor Virality Scoring: Evaluates candidate segments on a 0–100 viral score assessing hook strength, viewer retention curves, topic cohesion, speech velocity, dramatic pauses, and punchlines.
• Smart Transcript Chunking: Efficiently processes hour-long transcripts through token-aware sliding windows (up to 7,500 transcript tokens per chunk with 60-second temporal overlap) to preserve narrative context without truncation.
• AI Providers Settings & Validation: Added a dedicated AI Providers configuration tab with instant API key validation, model selection, and live connection testing.
• Onboarding Setup Step: Added an intuitive Gemini AI setup screen to the initial onboarding wizard, allowing users to configure and test AI clipping credentials before starting their first project.
• Configurable Clip Generation: Tailor clip generation parameters including target duration, minimum/maximum lengths, maximum clip count, and viral sensitivity thresholds.

Interactive Captions & Transcript Editor

• In-App Cue & Word Editor: Integrated a visual transcript editor directly inside the Results panel, allowing creators to edit text, correct misspellings, adjust word timings, and refine sentence boundaries.
• Interactive Player Seeking: Clicking any cue or word in the editor instantly seeks the video player to that exact timestamp and begins playback.
• Real-Time Dynamic Re-Burning: Saving transcript changes automatically triggers seamless background re-rendering, instantly updating the video preview without requiring a full project restart.
• Synchronized Playback Highlighting: Real-time active cue highlighting and smooth auto-scrolling keep the transcript synchronized with the video playback head.
• Optional Captions Toggle: Added the ability to completely disable captions generation and rendering for cutout-only or raw short-clip workflows.

Advanced Face Tracking & 9:16 Split-Screen Engine

• Upgraded Tracking Engine & ONNX Detection: Integrated an advanced face tracking pipeline powered by OpenShorts and ONNX-runtime face detection with mouth-motion lip tracking for active speaker detection.
• Dynamic Dual-Speaker Split Layout: Automatically generates stacked dual-speaker compositions (speaker 1 top, speaker 2 bottom) tailored for podcast interviews and dialogue scenes.
• Aspect Ratio Safeguard: Automatically disables split-screen layout on aspect ratios other than 9:16 (such as 1:1, 16:9, and 4:5) to prevent distorted video cropping.
• Framing Stability & Smoothing: Eliminated rapid camera jitter and jumping between filled and contained states using enhanced IoU tracking, temporal smoothing, and hysteresis thresholds.

Enhanced Project Management & Workspace UX

• Automatic Project Thumbnails: Automatically extracts and displays video keyframe thumbnails for all projects in the workspace grid.
• Complete Disk Cleanup on Delete: Deleting a project now cleans up all associated media files, extracted audio tracks, and temporary caches from disk.
• Folder Rename Synchronization: Renaming a project within the UI automatically updates the underlying directory on disk and atomically updates database storage references.
• Project Card Context Menu: Quick-action context menus on project cards for instant opening, renaming, and project deletion.

Fixes & Improvements

• Windows Face Tracking Stability: Offloaded face detection and tracking processing to non-blocking worker threads, eliminating "Not Responding" UI freezes on Windows during generation.
• Intel Windows Compression Fix: Tuned FFmpeg hardware acceleration and encoding parameters to resolve slow video compression speeds on Windows systems with Intel graphics.
• Rendering Pipeline Acceleration: Optimized caption overlay compositing and frame rendering routines for faster clip export times.
• Deepgram API Network Stability: Hardened cloud transcription network calls with expanded upload timeouts, retry logic, and resilient error recovery.
• Results Tab Player Stability: Fixed an issue where the video preview player could freeze or reset playback position when scrubbing through generation progress states.
• Urdu & Hindi Typography Support: Increased caption line height and improved glyph bounding box calculations to ensure clear rendering and prevent clipping of Arabic, Persian, and Devanagari descenders.
• Accurate Caption Previews: Fixed an issue where the style preview displayed arbitrary placeholder words instead of actual video transcript text.
• Preset Styles & Animation Refinements: Reorganized style control cards, improved color pickers, and enabled default animations for newly created projects.
• Actionable Error Messages: Implemented human-readable, context-specific error banners across transcription, AI clipping, and video export pipelines.
• Settings Modal Validation: Fixed an issue where the Settings dialog could be closed when required AI provider model fields were missing.

FuseClip 0.71.0

FuseClip v0.71.0

FuseClip v0.71.0 brings intelligent active speaker face tracking with mouth-motion detection, customizable text stroke and outline styling for captions, hardware-aware transcription model recommendations, a redesigned onboarding experience, native disk-backed project persistence, and major performance optimizations across the local Whisper engine and video pipeline.

What's New

Active Speaker Detection & Stable Face Tracking

• Mouth Motion & Lip Tracking: Integrated MediaPipe Face Landmarker with a dedicated MouthMotionTracker that analyzes jaw openness and speech dynamics in real time, automatically focusing the camera crop on whoever is actively speaking in multi-person scenes.
• Temporal Face Locking (PrimaryFaceTracker): Eliminated camera jitter and sudden jumping between faces using IoU tracking, candidate hysteresis, and switch margins.
• Adaptive Frame Sampling: Dynamically scales frame sampling rates for long videos (up to 1+ hour files) to keep face tracking fast and responsive on any CPU or GPU.

Caption Stroke & Outline Customization

• Custom Stroke Controls: Added configurable stroke width (0–30%) and stroke color pickers for single-word / karaoke highlights, preset styles, and sticker captions.
• High-Contrast Readability: Ensures burned and previewed captions remain crisp and legible against bright, high-contrast, or fast-moving video backgrounds.

Smart Model Recommendations & Guided Tooltips

• Hardware-Aware Suggestions: Evaluates system specifications (Apple Silicon Metal GPU, NVIDIA CUDA, CPU cores, RAM) to recommend optimal Whisper models (such as Large v3 Turbo) for maximum transcription accuracy without slowdowns.
• Language-Specific Deepgram Recommendations: Contextual tooltips detect selected languages and recommend optimal cloud models (e.g. Deepgram Nova-3 for multilingual code-switching, Urdu, Arabic, and Hinglish) with one-click model switching.
• Forced Language Selection: Requires explicit language selection in the Transcribe panel to ensure accurate transcription and transliteration settings before generation begins.

Redesigned Onboarding Experience

• Interactive Hardware Check: Updated Get Started screen with automated system diagnostics and clear side-by-side guidance between local Whisper and Deepgram cloud transcription.
• Inline API Key Validation: Added real-time Deepgram credential testing directly within the onboarding setup.

Fixes & Improvements

• Modular Whisper.cpp Engine: Refactored the local transcription service into dedicated, maintainable modules (runner, token-parser, downloader, binary-resolver, backend).
• Precision DTW Timestamps: Implemented Dynamic Time Warping (--dtw) token parsing with fallback heuristics to deliver precise, non-overlapping word-level timing and natural punctuation placement without artificial pauses.
• Whisper Performance Optimization: Resolved sluggish local model inference speeds by tuning execution parameters, thread allocation, and library loading paths.
• Native Disk-Backed Project Store: Migrated project persistence from localStorage to atomic file writes on disk (<userData>/projects/) via Electron IPC with automatic migration for existing projects.
• Extended Audio Extraction Timeout: Increased FFmpeg audio extraction timeout from 30 seconds to 10 minutes with an expanded buffer to handle long, high-resolution source media without timeouts.
• Pipeline Cancellation State Cleanup: Fixed an issue where cancelling a generation run and restarting with fewer clips would re-process stale clips from the previous canceled job.
• Interactive Generation Pipeline Progress: Added a dedicated preparation step to the generation ladder to clearly indicate disk read and model initialization progress.
• Tooltip Persistence: Fixed an issue where recommendation tooltips would not reappear after switching between local models and cloud providers.

FuseClip 0.7.0

FuseClip v0.7.0

FuseClip v0.7.0 introduces Deepgram cloud transcription support with live credential validation, intelligent audio quality diagnostics and warnings, automatic GPU acceleration detection, built-in phonetic Hindi and Urdu transliteration, and bundled Windows Visual C++ runtimes for seamless out-of-the-box local transcription.

What's New

Deepgram Cloud Transcription

• Cloud Provider Integration: Added native integration with the Deepgram API as an alternative to local Whisper models, offering lightning-fast, high-accuracy transcription.
• Smart Formatting & Features: Automatic language detection, smart punctuation, formatting, and utterance splitting out of the box.
• Live Credential Validation: In-app API key management in Settings with real-time connection and credential verification via IPC.
• Flexible Provider Switching: Easily switch between Local (Whisper) and Cloud (Deepgram) modes right inside the Transcribe panel and Settings.

Audio Quality Diagnostics & Warnings

• Automated Audio Diagnostics: Pre-transcription audio analysis powered by FFmpeg detects potential audio quality issues before transcribing.
• Actionable Warning Tooltips: Real-time warnings for low audio levels, clipping/distortion, high background noise, and fast speech rate, complete with suggestions on how to achieve optimal transcription results.

GPU Hardware Detection & Recommendations

• Automatic GPU Detection: Detects system hardware capabilities including Apple Silicon Metal and NVIDIA CUDA acceleration.
• Hardware Recommendations: Provides hardware-tailored local Whisper model recommendations based on available compute and VRAM.

Hindi & Urdu Transliteration Engine

• Phonetic Transliteration: Built-in phonetic transliteration engine for converting Indic (Devanagari) and Nastaliq (Urdu) transcripts into Roman Urdu and Hinglish captions.
• Accurate Timing & Segmenting: Preserves word-level timing precision and caption boundaries during transliteration.

Fixes & Improvements

• Bundled Windows VC++ Runtime: Packaged Microsoft Visual C++ redistributable binaries (vcruntime140.dll, msvcp140.dll, vcomp140.dll) directly into the Windows app packages, resolving missing DLL errors on clean Windows installations.
• Workflow State Preservation: Fixed an issue where transcription results could be reset during step transitions or project navigation.
• Deepgram IPC Validation: Resolved renderer-side network validation issues by routing Deepgram validation requests through main process IPC.
• Transcription Timing & Segmentation: Fixed edge cases in word timestamp alignment and sentence segment boundaries.
• Windows Auto-Updater Elevation: Hardened UAC elevation handling during background updates on Windows.
• UI & Icon Modernization: Updated UI components with modern icon sets, refined dialog animations, improved tooltips, and responsive layout adjustments across the sidebar and panels.

FuseClip 0.6.0

FuseClip v0.6.0

This release introduces dynamic caption entrance animations with spring physics, per-project style memory, major transliteration upgrades for Hinglish and Roman Urdu, local Whisper model management in settings, FuseGrab video link integration, and UI refinements.

What's New

Caption Entrance Animations

• New Animation Tab: Choose from multiple entrance animations including Pop, Fade, Slide Up, Slide Down, Slide Left, and Slide Right (or None).
• Fine-Tuned Motion Controls:
  • Speed: Adjust animation playback rate from 0.1× to 3×.
  • Distance: Control the slide travel distance from 0.1× to 4×.
  • Bounce: Dial in springy overshoot and oscillation dynamics (from smooth ease-out to high-energy bounce).
  • Pop Start Scale: Configure the initial scale (0%–90%) for the pop entrance effect.
• Word & Block-Level Entrances: Seamless support for animating whole caption cues together or animating individual words consecutively in Single Word display mode.
• Synchronized Live Preview & Export: Frame-accurate animation rendering across both the real-time canvas preview and the final video burning process.
• Updated Presets: Core style presets now come pre-configured with complementary entrance animations.

Per-Project Preset Memory

• Project-Scoped Working Styles: Modifications made to preset styles are now saved per-project. Switching between presets retains your custom tweaks within that project without losing progress.
• Clean Project Defaults: New projects always start fresh with pristine default preset styles.

Whisper Model Management & Recommendations

• Downloaded Models Dashboard: View all cached local Whisper models along with their exact sizes on disk in Settings.
• One-Click Model Deletion: Easily remove downloaded GGML models from disk to reclaim storage space.
• Smart Model Recommendations: Contextual warnings and one-click model suggestions when a selected model size may underperform for the chosen language.

FuseGrab Video Link Integration

• Direct Video Downloader Access: Added a quick shortcut in the video selector linking to FuseGrab, allowing users to download videos directly from URLs before importing them into FuseClip.

Fixes & Improvements

• Results Panel Polish: Cleaned up clip listing layouts and score badge displays across both viral-clips and captions-only workflow modes.
• Inactive Opacity Support: Whole-line opacity slider now properly applies when word-level highlights are turned off.
• UI & Component Refinements: Fixed notification badge ping styling in the sidebar and perfected tooltip arrow alignment.

FuseClip 0.5.0

FuseClip v0.5.0

This release focuses on fixing caption timing accuracy and word synchronization.

Fixes & Improvements

Accurate Caption & Word-Level Timing

• Dynamic Time Warping (DTW) alignment: Transcription now uses whisper.cpp's DTW token timestamps (--dtw) instead of default heuristic timestamps, eliminating timing drift where word highlights lagged behind or drifted from spoken audio.
• Precise word boundaries: Word boundaries are now resolved against the next token's onset, ensuring each word has an accurate start and end timestamp.
• Snappier visual highlight: Word highlights lead audio onset slightly (80ms) to synchronize naturally with spoken speech and visual perception.
• Graceful fallback: Automatically falls back to standard timestamps if an installed model or binary does not support DTW presets.