Video Candy

Enables frame-accurate video trimming directly in the browser using WebGL-accelerated canvas rendering and client-side video codec libraries (likely FFmpeg.wasm). Users set in/out points on a timeline scrubber, and the tool generates a new video file without server-side processing for files under size limits, reducing latency and privacy exposure compared to cloud-based editors.

Provides pre-designed video templates optimized for TikTok (9:16), Instagram Reels (9:16), YouTube Shorts (9:16), and landscape formats (16:9) with built-in text overlays, transitions, and music placeholders. Templates are stored as JSON-serialized composition graphs that map media layers, timing, and effects, allowing users to drag-and-drop content into predefined slots without manual layout work.

Embeds a Video Candy watermark (logo and text) into the bottom-right corner of exported videos on the free tier. The watermark is rendered as a PNG overlay during export using FFmpeg's overlay filter, positioned at a fixed location with configurable opacity (50-100%). Premium users can disable the watermark or replace it with custom branding (logo image and text).

Provides a curated library of 50+ pre-built transitions (fade, slide, zoom, blur) and visual effects (color overlay, brightness adjustment, blur) implemented as WebGL shaders. Users select a transition type and duration (0.3-2 seconds), and the tool automatically generates the intermediate frames by interpolating between source and destination video frames using GPU-accelerated blending.

Exports edited videos to MP4, WebM, and MOV formats with automatic bitrate optimization based on target platform (TikTok: 2.5-4 Mbps, Instagram: 3-6 Mbps, YouTube: 5-15 Mbps). The export pipeline uses FFmpeg with preset encoding profiles that balance file size and quality, and applies platform-specific metadata (aspect ratio, duration limits) to ensure compliance with platform requirements.

Allows users to adjust volume levels for video audio tracks and add royalty-free background music from an integrated library using a simple slider interface. The audio mixing is performed at export time using FFmpeg's audio filter graph, which combines the original video audio and background music tracks with specified volume levels (0-100%) and applies basic crossfading between tracks.

Enables users to add text overlays and captions to video frames using a text editor that applies preset styling templates (bold, italic, shadow, outline). Text is rendered as a separate layer in the composition graph with configurable duration, position (9-point grid), font size, and color. The text rendering uses Canvas 2D text rendering at export time, with automatic font fallback for unsupported characters.

Converts videos between aspect ratios (16:9, 9:16, 1:1, 4:3) by either letterboxing (adding black bars), pillarboxing (adding side bars), or cropping to fill the target frame. The conversion is performed at export time using FFmpeg's scale and pad filters, which resize the source video and add padding with configurable background color, or crop to the target dimensions.

Provides a horizontal timeline interface where users arrange video clips, images, and audio tracks in sequence with frame-accurate positioning. The timeline uses a layer-based composition model where each clip occupies a track, and clips can overlap to create picture-in-picture effects. The timeline state is stored as a JSON composition graph that maps clip positions, durations, and layer order, enabling undo/redo and session persistence.

Allows users to import video, image, and audio files via drag-and-drop onto the canvas or through a file picker dialog. Imported files are stored in browser IndexedDB with metadata (duration, dimensions, codec) extracted using FFmpeg.wasm. The asset library displays thumbnails and allows users to organize files into folders and search by filename, with automatic deduplication based on file hash.

Provides a video preview window that renders the current timeline composition in real-time as users make edits, with playback controls (play, pause, seek) and frame-by-frame scrubbing. The preview uses Canvas 2D or WebGL to composite layers from the timeline and render the result, with frame rate capped at 30fps to balance responsiveness and browser performance. Scrubbing seeks to specific frames without full playback.