Qamrec - Private Screen Recorder & Subtitles

Makes recordings look edited: auto-zoom on clicks, cursor and keystroke effects, webcam styling, subtitles. All on your device.

Qamrec - Private Screen Recorder & Subtitles A free, open-source, privacy-focused screen and camera recorder for Chrome that makes recordings look edited: automatic zoom on clicks, smooth camera moves, cursor and keystroke overlays, a styled webcam bubble, on-device subtitles, and a built-in editor. Everything is processed on your device. Features Recording • This tab - One click, no picker; enables all page effects (auto-zoom, clicks, keystrokes, typing focus) • Screen / window / tab - Chrome's picker; page effects apply when the shared surface is the browser window you started from, otherwise smart auto-framing is used • Camera only, or a webcam overlay recorded as a separate track so it can be restyled after recording • System/tab audio + microphone, mixed into one track, with noise reduction and live level meters • Pause & resume, live preview, countdown, highlight markers, no artificial time limit • Scheduled recording (start in N minutes or at a time) and auto-stop • Tutorial mode - numbered click steps, keystrokes and cursor highlight • Presets (Cinematic, Tutorial, Clean capture, Quick GIF, Talking head, plus your own) • Keyboard shortcuts: Alt+Shift+S stop, Alt+Shift+P pause/resume, Alt+Shift+M marker Cinematic effects (applied in the editor, fully adjustable) • Automatic click zoom with smooth zoom & pan that anticipates the cursor ("follow the action") • Automatic typing focus and element focus glow on the field being typed into • Smart auto-framing from on-screen motion when page tracking isn't available (e.g. desktop apps) • Cursor highlight, click ripples, numbered click indicators, scroll indicator • Keystroke & shortcut display (typed text is never shown; password/payment fields are ignored) • Custom zoom intensity, hold duration and transition speed; manual zoom keyframes (add, move, resize, set focus by clicking the preview) • Delete individual auto zooms you don't want, and restore them all again later • Webcam shape (circle, rounded, square, wide), position (drag it), size, crop, border, shadow, mirror and background blur (MediaPipe, runs locally) • Frame styling - background gradients, padding, rounded corners, drop shadow • Chapter markers from page changes and scene cuts (with optional title cards) and automatic highlight detection Subtitles • Generate subtitles from the audio on your device with OpenAI's open-source Whisper model (GPU via WebGPU when available, otherwise CPU/WASM); choose the language or detect it, and Fast / Balanced / Accurate model sizes • Word-level timing splits speech into readable cues at pauses and sentence ends; silent stretches and cut parts are skipped • Edit and manage: fix text, add at the playhead, delete, split, merge, set start/end to the playhead, drag or resize cues on the timeline, shift all cues to fix sync • Style: burned into the video (position, size, color, background box or outline, characters per line), plus an optional embedded WebVTT track that viewers can toggle (WebM only - the MP4 container cannot carry one) • Import / export SRT and VTT (exports follow your cuts and trims) Editing & export • Trim & cut on a timeline with waveform, automatic silence removal, "keep only highlights" • MP4 (H.264/AAC via WebCodecs, falling back to HEVC/VP9/AV1 if the machine can't encode H.264), WebM (VP9/Opus) and GIF export, with conversion progress and cancel; GIFs are capped at 720p / 15 fps and carry no audio • Resolution scaling (original, 2160p-480p; scales by height and never upscales) and frame-rate choice • Custom filename templates: {date} {time} {datetime} {title} {mode} {duration} {res} {n} • Save the original recording (WebM, no effects applied) alongside the edited export • Save the current look as a preset from the editor, to reuse on later recordings • Settings are grouped into Zoom, Cursor, Camera, Frame, Edit, Subtitles and Export panels Privacy All recording, analysis, transcription and encoding happens locally in your browser. Nothing is uploaded, there is no account, no analytics. The only network request is the optional, one-time download of the open Whisper model weights from Hugging Face when you first generate subtitles (then cached); your audio never leaves your device. The page tracker is injected only into the tab you start recording from, only sends events while recording, and never records typed characters.