Introducing CstWeave: A 100% On-Device, Privacy-First Tab Region GIF & WebM Recorder
Hey everyone! 👋
I’m excited to share CstWeave , a privacy-first Chrome extension I built for capturing and exporting tab-region recordings directly into GIFs or WebM videos —100% on-device.
Why CstWeave?
Most screen recorders stream frame data to external cloud servers, mandate account sign-ups, enforce render queues, or slap forced watermarks on free exports.
When you're recording internal admin dashboards, logging bug reproductions, or capturing client tools containing sensitive credentials, uploading media to third-party cloud servers is a major security risk. CstWeave keeps all frame buffers strictly inside your browser memory.
Key Highlights
100% On-Device Processing: Capturing, cropping, and encoding run entirely inside local RAM/CPU. No media touches a remote server or cloud queue.
Sub-second Hotkey Trigger: Press
Alt+Shift+G(⌘+Shift+GorOption+Shift+Gon Mac) on any tab to instantly select a region and record.No Forced Watermarks: Enjoy clean, pristine exports on both free (Lite) and paid (Pro) tiers.
Custom Pure-TypeScript GIF Engine: Built with an offscreen pipeline using Median-Cut palette sampling, Floyd-Steinberg dithering, and LZW encoding to avoid background frame throttling.
Pay Once, Use Lifetime: Lite is free forever ($0); Pro is a one-time $9.99 upgrade (not a recurring subscription) for HD WebM, longer recording limits, and audio capture.
Tech Stack
Extension Architecture: Built for Chrome Extension Manifest V3 using TypeScript, Shadow DOM overlay (for scoped UI), Service Worker, and the Offscreen Document API.
Video/Audio Pipeline: Web
MediaRecorderAPI overcanvas.captureStream()with optional system audio track mixing.
I’d love to get your feedback, hear your thoughts on the custom GIF pipeline, or answer any technical questions! Check it out at https://cstweave.com.
Love the on-device choice here. That's the harder engineering path, not the easier one.
On Aryan's question: privacy-via-browser is underrated because most users don't shop for it consciously. They wake up not wanting to get fired for uploading a client dashboard to a random cloud queue, not asking for "on-device processing" by name. The privacy is what lets them say yes at all. The speed is what they'll actually tell you they like.
Bigger question though: if a browser sandbox is trustworthy enough to keep frame buffers local for this, why do we accept AI agents that can't say the same about our files? I'm testing that exact theory with something similar in spirit, an agent that works on real local files entirely inside the browser tab, nothing uploaded. Same reaction every time: "wait, this actually stays on my machine?"
Are you seeing that reaction too, or does it matter less for GIFs and more just about no watermark?
The privacy-first angle is strong, especially for recordings involving sensitive dashboards or client data.
The interesting question is whether users mainly switch because of privacy, or because the faster local workflow removes enough friction from existing screen recorders.
“Privacy-first is a great angle. Curious — what’s the moment where users realise they need this vs it just being nice to have?”