Changelog

• Screenshot Mode — Capture and edit screenshots with a dedicated editor. Includes shapes, text annotations, sensitive masking, backgrounds, padding, and more. Export or share with a single click.
• Local caption generation — Generate captions directly on your device using Whisper, with support for many languages. Fully customizable styling (font, size, color, background, position) and SRT export.
• New editor track types — Text overlays, Sensitive Mask, and Highlight tracks give you more creative control over your recordings.
• Overhauled recording picker — Brand new target select overlay makes it easier than ever to pick your recording source.
• Recoverable recordings for Studio Mode — Recordings are now saved in 3-second chunks. If anything goes wrong, you'll lose at most 3 seconds—full recovery system built in and enabled by default.
• Circle cursor — New mobile simulator cursor option for showcasing app demos.
• Faster editor playback — Major performance improvements including WebGPU rendering, optimized YUV conversion, and hardware decoder support on Windows. Longer videos now scrub and play back much smoother.
• Preview quality options — Quarter and Half preview quality options for faster playback on slower systems.
• Improved audio-video sync — Post-recording sync calibration, drift correction, and input latency estimation on macOS for tighter sync.
• Recording hotkeys — New pause/resume hotkey and cycle recording mode hotkey. Tray menu now shows mode selection with dynamic icons.
• Better device support — Lots more webcams and microphones are now supported, with improved device detection and permission flows.
• Redesigned export page — New export UI with live preview, render time estimates, and frame-accurate cropping via a timeline slider.
• Windows reliability — Hardware compatibility checks, encoder health monitoring with automatic fallback, Media Foundation decoder with D3D11 interop, and Windows fragmented M4S muxer support.
• Bug fixes — Fixed memory leak while streaming, split action now respects segment speed, and recordings now use human-readable project names instead of UUIDs + lots more quality of life improvements.

• New Instant Mode Resolution Picker — You can now set the resolution of Instant Mode recordings in Settings. Choose from 720p, 1080, 1440p and 4k. Defaults to 1080p.
• Uploader speed improvements — Uploads are now done more efficiently by uploading multiple chunks at the same time, bumping the chunk size to 20mb, and reusing the presigned URL for the last chunk
• Audio recording fixes — A bug that caused recorded audio to sound choppy should now be fixed
• Upload logs feedback — Using the 'Upload Logs' tray button will show a dialog if the logs upload succeeds
• Improved area selector — Drag to select now supported in the new recording flow area selector (enable new recording flow via Settings -> Experimental)
• Improved error handling — Some errors that can occur while recording on Windows will now safely stop recording

• Quicker exports — export pipeline optimizations cut total export time by ~16%
• macOS Tahoe support — high-quality SVG Tahoe cursor icons, and fresh macOS Tahoe wallpapers
• Bluetooth playback fixed — improved audio latency handling & lag on wireless headphones
• Fixed audio device selection — better mic device detection (e.g., AirPods now appear reliably)
• Snappier uploads — optimistic presigned URLs reduce upload latency
• New uploader — the new uploader is now on by default for Instant Mode, increasing upload reliability
• Safer pausing — enforce monotonic PTS for audio/video to prevent corruption after pauses in Instant Mode
• Zoom workflow upgrades — split Zoom and Scene segments, fixed an issue with clicking zooms segments on long recordings, and you can now drag to create new zoom segments;
• Multi-monitor improvements — tighter Target Select overlay bounds and corrected List View in the new recording picker (enable via Settings > Experimental)
• Stabler capture/encode — don't abort if the capturer stalls, proper DXVA2 hardware decode on Windows, and handle "no frames" from screen capture
• Hide inactive cursor — you can now set cursor idle time in the editor, automatically hiding the cursor after X amount of seconds with a smooth fade
• Editor UX touches — scissors cursor shows when the Clip tool is active
• Quality of life — On Windows, the main window now re-opens after closing Settings/Editor

• New Window Excluder — hide specific apps/windows from recording (works best on macOS, enable via Settings → General)
• Camera & mic safety — we now hold device locks only while recording is active, preventing “device busy” conflicts and improving handoff between sessions
• Multi-channel USB audio — proper support for interfaces with multiple input channels
• More faithful color — AVFoundation recordings now use the Rec.709 color space for better color accuracy on modern displays
• Stronger uploads — the new uploader is enabled on desktop, and we’ve added multipart uploads for Studio recordings and Instant re-uploads to handle large files and spotty networks
• Corruption guard — the muxer now stops cleanly if a frame queue failure occurs, avoiding partial or corrupted outputs
• Clearer errors — improved messaging and handling when video upload info can’t be found
• Under-the-hood reliability — numerous recording pipeline hardening tweaks for steadier performance

• New List View in the recording picker (enable via Settings → Experimental).
• Faster recording start on macOS by sharing ScreenCaptureKit warm-up across sessions
• Windows capturer now runs on a dedicated thread for snappier, more reliable capture
• Windows encoding is more resilient: automatic fallback to software H.264 when hardware ProcessInput fails, plus fixes to the software path
• Auto-zoom feels smoother on fast cursor moves with refined segment generation and steadier cursor behavior (including safer cursor-actor cleanup)
• Fixed a case where the 5-minute Studio export limit could be enforced incorrectly
• Improved app icon/logo render quality in the new recording picker on Windows
• Added minidump crash reporting (Sentry via Tauri) to help diagnose rare crashes.
• Clearer “needs input” error handling in the desktop flow.
• Improved recording pipeline under the hood for improved reliability and lower latency

• Complete camera system overhaul, enabling broader device support, lower latency and smoother performance
• Experimental: Auto zoom segments automatically inserts new zoom segments at key moments.
• Cursors now render as high-resolution SVGs for better quality (when available)
• Configurable countdown before recording starts (3, 5 or 10 seconds)
• Export dialog now shows exact output resolution before you render
• Recording halts if sustained frame drops are detected (so you are aware earlier that the recording has stopped)
• On Windows, launching Cap again jumps straight to the recorder instead of opening a second instance
• Refreshed settings pages design
• Edit the name of a recording directly in the editor

Enable "Experimental" features via Settings > Experimental.