Decentralizing video processing to the edge of your browser.
FrameExtract was built on a simple conviction: you shouldn't have to upload sensitive multi-gigabyte video files to remote servers just to extract frames, trim clips, or rip audio.
Why Client-Side?
Traditional online video convertors require uploading full video files to cloud backends. This introduces enormous bottlenecks: slow upload speeds, server bandwidth caps, monthly subscription paywalls, and critical privacy vulnerabilities for sensitive footage.
With the arrival of standard browser capabilities like the WebCodecs API, OffscreenCanvas, and the Origin Private File System (OPFS), modern browsers now have direct access to your local GPU hardware decoders. FrameExtract unlocks this power with zero server overhead.
100% Zero-Server
Your video bytes never leave your device. All decoding and rendering happens in sandboxed Web Workers.
GPU-Accelerated
Native WebCodecs hardware decoders achieve extraction speeds of 100+ frames per second on modern machines.
Zero-RAM Safety
Streaming fflate architecture flushes frames to disk immediately, preventing browser memory crashes on long clips.
The Two-Tier Architecture
Tier 1: Native Engine (Instant 0KB WASM)
Used for Frame Extraction, Audio Extraction, and Lossless Trimming. Uses native browser APIs (VideoDecoder, mp4box.js, and fflate) for instant page loads.
Tier 2: WASM Engine (On-Demand)
Used for complex re-encoding like GIF palette creation, MP4 CRF compression, and hardcoded subtitle burning. Dynamically loads @ffmpeg/ffmpeg only when selected.
Open Source Foundations
FrameExtract stands on the shoulders of phenomenal open source projects: