Simple Web Recorder keeps recordings on your device.

What stays local

• Captured tab, screen, camera, and microphone media.
• Smart Voice processing, Voice Ducking noise-floor estimates, and temporary adaptive gate/voice-active levels; the native Web Audio graph runs only in the offscreen document.
• When explicitly enabled, a short-lived on-device live-caption phrase can move the Studio-only teleprompter after a conservative local script match. It is not added to the recording, retained as a teleprompter preview, or sent to a service.
• One user-selected local music bed (up to 25 MB), retained only in the extension database and decoded only for its current recording mix.
• Optional raw tab-audio and microphone WebM/Opus sidecars, only when you enable separate audio stems; they stay in the extension database for local editing/download.
• Incremental recording chunks and recovery information.
• Scene layouts, device preferences, and recording settings, including local five-band microphone EQ gains and narration level-assist settings but never microphone samples.
• Brand-kit colors, fonts, and optional locally resized logo images.
• Teleprompter settings, named scripts, and imported local text.
• Person-segmentation masks, including the current temporary matte for an opted-in local video clip or static image; temporary face landmarks and bounded Auto Tone/gaze values for an opted-in local clip portrait effect; temporary expression blendshapes, hand landmarks, canned gesture scores, temporary Privacy Shield face boxes, and caption transcripts.
• Temporary AI Smart Slim geometry, AI Auto Tone color measurements, AI Makeup Studio face regions, and local portrait settings.
• When you explicitly enable Local Color Focus, Chroma Key Studio, or Local Graphic Transform Lab, one temporary capped Canvas copy of the active Capture, Camera, or Image layer. Chroma Key may change alpha only for creator-selected screen colors; Graphic Transform Pixelate may average alpha inside its selected local blocks; ASCII samples that temporary copy into a transparent glyph grid; VHS creates one short-lived neighbor-sampling pixel copy that is zeroed after its pass. The copy is cleared immediately after each composition pass and is never uploaded or persisted.
• Camera Stabilizer Studio's one overwritten 160 px grayscale analysis frame, temporary full-resolution transform scratch, and bounded numeric motion diagnostics.
• Local AI Group Frame's overwritten 384 px-or-smaller detector input, temporary face boxes, and bounded numeric crop/response state.
• Local AI Framing Zone and AI Smart Punch-In's temporary packaged face landmarks plus bounded numeric crop target, engagement, and response state. Source camera frames are never stored.
• Local AI Background Parallax's temporary packaged face landmarks and bounded per-layer replacement-image x/y offsets.
• Local AI Cutout Shadow Studio's current temporary person-matte alpha and selected compositing controls; it creates no extra history.
• Local AI Presenter Anchor's temporary face bounds or detected-hand coordinates plus local Left/Right handedness label and bounded in-memory overlay position smoothing.
• Local Smart Presenter Director's current normalized source-tab pointer coordinate and bounded in-memory camera-position smoothing while explicitly enabled.
• Camera Denoise Studio's temporary 320 px scratch pixels, one overwritten downscaled derived history frame, and bounded numeric noise/motion diagnostics.
• Local Camera Flicker Guard's overwritten 96 px global-luminance sample, temporary full-resolution brightness scratch, and bounded numeric baseline/correction state.
• Local AI Clip Flicker Guard's overwritten 96 px global-luminance sample, temporary full-resolution brightness scratch, and bounded numeric baseline/correction state for a muted local clip.
• Local AI Adaptive Grade's overwritten 192 px luma sample, temporary full-resolution filter scratch, and four bounded numeric response values for a camera or muted local clip.
• Local AI White Balance's overwritten 192 px RGB sample, temporary full-frame pixel buffer, and four bounded numeric response values for a camera or muted local clip.
• One creator-selected Local AI Color Match PNG/JPEG/WebP reference (up to 2 MB) in local scene settings, overwritten 192 px source/reference samples, temporary full-frame pixel buffer, and bounded numeric response values for a camera or muted local clip.
• Local procedural motion-background colors, direction, depth, and animation time.
• Temporary Smart Highlight scores and editable clip ranges derived from local captions and recording markers.
• Temporary Local Retake Finder suggestions derived only from final local caption text and cue times.
• Editable local Focus Track timing and normalized geometry, including creator-accepted caption punch-ins; it never stores caption text in a focus window.
• Existing creator-approved local marker times and labels are read only while a requested post-production Chapter Card export is encoding; Chapter Card style, duration, and accent remain page controls and create no new analysis.
• Temporary local Caption Search & Find query, result, and preview-navigation state.
• Temporary local Caption Cut selections and editable cue-time ranges.
• Temporary local Visual Change Finder low-resolution color/luminance signatures and candidate times.
• One temporary same-dimension Canvas only while encoding a user-requested local PNG frame.
• One temporary bounded contact-sheet Canvas only while encoding a user-requested local PNG storyboard.
• One temporary local thumbnail Canvas only while encoding a creator-requested PNG thumbnail.
• A temporary bounded failed-shortcut command name, error message, and timestamp in browser-session storage.
• Post-production caption styles, estimated active-word state, and locally rendered export pixels.
• Four numeric Local Capture Area edge-crop settings saved with the local scene; they change only the composed output, not Chrome's authorized source.
• One local Capture Presentation Frame style and optional user-entered browser label saved with the scene; no page title, URL, asset, or media is read to create it.
• Normalized cursor movement and click coordinates used for local recording effects. Optional Cursor Momentum Blur keeps only the immediately previous normalized point while drawing a visible cursor, then clears it on leave, disable, or stop.
• When Local Sensitive Field Shield is explicitly enabled, only the temporary normalized on-screen rectangle of the focused password or browser-standard payment input. No value, label, field name, ID, or page text is read.
• When Local Privacy Curtain is explicitly enabled, only the temporary true/false state of the explicit Alt+Shift+B hold in the focused captured tab. No page content, form value, URL, or general keystroke is read or sent.
• When Local AI Screen Face Shield is explicitly enabled, temporary capped-resolution Capture pixels are processed by the packaged on-device face detector. It clears each input after inference, never saves a frame or face box, and uses only short-lived in-memory numeric boxes for smoothing—never identity or page text.
• A local Push-to-Talk enabled setting and the temporary true/false state of the explicit Alt+Shift+U hold; no page text, DOM content, URL, form value, or other keystroke is read or sent.
• Temporary Local Microphone Readiness Check RMS/peak values while you explicitly test a selected input; samples and classifications stay in the setup page and clear when the test stops.
• The title, URL, and favicon URL of a tab you explicitly start recording, stored only as local session and recording metadata so the Hub and Library can identify that source.
• Finished WebM files until you download or delete them.
• If you separately install Simple Capture Bridge and explicitly enable Send to OBS, local pairing, feed-control, WebRTC signaling, and short-lived session tokens travel only through Chrome Native Messaging to the registered com.rippax.simple_capture_bridge companion on this device. Selected live media is mirrored separately over WebRTC to a receiver bound only to 127.0.0.1; no cloud relay or remote ICE server is used.
• Visible-tab and full-page screenshots, including annotator marks, PNG/PDF exports, and clipboard copies. They stay in Chrome Downloads or a session handoff on this device.

Why each permission is requested

Active tab

Identifies the tab you deliberately choose from the extension toolbar.

Scripting

Installs the cursor/Push-to-Talk/Sensitive Field Shield/Privacy Curtain helper only in the current tab you explicitly start recording. It observes enabled pointer events, the explicit Alt+Shift+U and Alt+Shift+B hold/releases, and (only when opted in) the focused password/payment field's rectangle—never page text, DOM content, field values, labels, names, IDs, or general keystrokes. The same permission injects a full-page screenshot scroller into the tab you chose from the toolbar; that helper reads only scroll positions and viewport size, never page text or DOM content.

Tab capture

Receives that tab’s video and audio. Chrome suppresses local tab playback during capture unless you explicitly enable monitoring.

Desktop capture

Shows Chrome’s native picker when you choose a window, browser tab, or full screen.

Offscreen document

Keeps media recording and any selected local duration auto-stop active after the popup closes and while the Manifest V3 service worker sleeps.

Side panel

Shows the local Recording Hub with each explicitly started tab's state, elapsed and remaining time, size, and controls. It does not authorize capture of another tab.

Storage

Saves settings and recording chunks locally for long-session safety and recovery.

Unlimited storage

Lifts Chrome’s default per-extension storage cap so long recordings are limited by your free disk space rather than an arbitrary quota. Everything still stays on this device; the recorder warns and stops safely as your disk fills.

Downloads

Lets you save a selected recording to your computer.

Desktop notifications (optional)

Requested only when you enable recording alerts in the Hub. Chrome may show a captured tab title and local start, pause, ready, error, or one-minute-left state; no media, page content, URL, or transcript is sent anywhere. Turning alerts off removes the optional permission.

Native Messaging (optional)

Requested only when you enable Send to OBS. It pairs with the separately installed com.rippax.simple_capture_bridge companion and exchanges local feed-control and WebRTC signaling messages; media never travels through Native Messaging. Turning Send to OBS off disconnects the companion and removes the permission.

Camera and microphone

Requested by Chrome only when you enable those sources.

What the extension does not do

• It does not upload recordings, URLs, titles, audio, video, AI frames, or transcripts.
• Send to OBS is off by default and does not silently start OBS recording or streaming. When enabled, its capability-bearing receiver URL is intended only for the user's local OBS Browser Source and should be treated like a local password while live. OBS stores that URL in the user's local scene collection so the source can reconnect; its token expires when the companion session exits. Local extension recording continues independently if the companion or OBS disconnects.
• It does not read or transmit page text, HTML, form values, or links for cursor enhancement.
• Full-page screenshot capture does not read page text, HTML, form values, or links. The injected helper only scrolls and reports scroll positions and viewport size for the tab you explicitly chose.
• Local Sensitive Field Shield reads no field value, label, name, ID, or page text. It uses only a focused password or browser-standard payment input's temporary rectangle to blur the locally composed fixed-tab recording, then clears that geometry on focus loss, setting change, navigation, or stop.
• Local Privacy Curtain reads no page content or general keystrokes. It acts only on the explicit Alt+Shift+B hold/release in a fixed current-tab recording, replaces the composed Capture pixels locally, and clears on release, focus loss, navigation, pause/stop, or setting change.
• Local AI Screen Face Shield detects no identity and reads no page text. Only while explicitly enabled, it uses a packaged local face detector on temporary capped-resolution Capture frames to blur or pixelate every confident face, then clears the input and never saves frames or face boxes; short-lived numeric boxes exist only in memory for smoothing.
• Push-to-Talk does not read or transmit general keystrokes. It uses only the explicit Alt+Shift+U hold/release to gate the recording microphone mix, closes on release/focus loss/navigation/pause, and never routes that mic mix to speakers.
• Local Microphone Readiness Check does not play, record, upload, or persist microphone audio or its temporary level/classification. It closes the test input when you stop it, close setup, change the device/mode, or start recording.
• Local Capture Area does not change Chrome's native picker selection or expand the source authorized for capture.
• Local Capture Presentation Frames do not download mockup assets, read tab titles or URLs, or alter the Chrome-authorized source.
• It does not monitor microphone audio through your speakers.
• It does not play a selected local music bed through your speakers, upload it, synchronize it, or send its bytes through extension messages.
• Separate audio stems are optional raw source tracks before the extension mix (no music bed, ducking, Push-to-Talk, or final limiter). They have no speaker route, are not uploaded, and are deleted with the recording or under storage pressure.
• It does not upload microphone audio, gate levels, narration level-assist level/gain decisions, Voice Ducking noise-floor estimates, voice-active state, or gain decisions for Smart Voice, Podcast warm, Broadcast punch, or Voice Ducking processing.
• It does not upload or retain expression scores used for AI reaction effects.
• AI gesture reactions and Local AI Air Pointer process temporary hand landmarks, index-fingertip coordinates, handedness, and canned gesture scores with a packaged model only while enabled. Pointer smoothing and trail points remain in memory. Those values are not uploaded, persisted, used for identity, or retained as biometric templates.
• Local AI Privacy Shield uses temporary face boxes and confidence scores from a packaged detector. It performs no identity recognition, enrollment, embeddings, or biometric matching, and does not upload or persist frames or boxes.
• Local AI Eye Contact Studio uses temporary packaged iris, eyelid, eye-corner, and face landmarks to deform only the original eye pixels. Source and patch scratch pixels are cleared immediately; only bounded numeric offsets are briefly smoothed in memory and cleared when the effect or camera stops. It does not upload or persist frames, patches, landmarks, or offsets, identify a person, retain biometric templates, or generate replacement eyes.
• Local AI Camera Stabilizer Studio uses one overwritten grayscale analysis frame capped at a 160 px long edge plus bounded numeric translation state. Its full-resolution transform scratch is cleared after every frame, and all state clears on disable, camera/settings changes, hide/removal, reset, or stop. It does not upload or persist pixels, samples, or motion diagnostics, identify a person, or synthesize replacement edge content.
• Local AI Group Frame uses temporary BlazeFace boxes from a packaged detector input capped at a 384 px long edge to crop around one or more people. The input canvas is cleared after every inference, and boxes/crop state clear on disable, camera changes, hide/removal, reset, or stop. It does not upload or persist pixels, boxes, or diagnostics; identify people; inspect audio; infer who is speaking; or synthesize image content.
• Local AI Framing Zone reuses only the latest temporary packaged face landmarks to aim Auto Frame's bounded one-face crop at Center, Left third, or Right third. AI Smart Punch-In uses only that current landmark availability to smoothly add a bounded crop boost, then release on face loss. Its numeric crop target and engagement state clear on disable, camera changes, hide/removal, reset, or stop; it does not upload or persist frames, landmarks, or crop state; identify a person; inspect audio; infer who is speaking; or generate pixels.
• Local AI Background Parallax reuses only the current packaged face-landmark center to move an opted-in replacement image opposite the presenter. It retains only bounded numeric x/y offsets and clears them on disable, image/source changes, hide/removal, reset, or stop. It does not upload or persist frames, landmarks, or offsets; identify a person; estimate depth; generate a 3D scene; inspect audio; or infer who is speaking.
• Local AI Cutout Shadow Studio reuses only the current temporary packaged person-matte alpha to draw a bounded neutral shadow beneath an opted-in segmented presenter. It creates no additional frame/matte history and does not upload or persist frames, mattes, or runtime values; identify a person; infer real lighting or depth; generate a new scene; reconstruct camera content; inspect audio; or infer who is speaking.
• Local AI Image Green Screen sends an opted-in creator-selected static image through the packaged local segmenter once per image/crop geometry, then retains only the current temporary confidence matte while the effect is active. The model input and matte close on source/crop change, disable, hide/removal, reset, or stop. It does not upload or persist image pixels, model input, masks, or runtime state; identify a person; reconstruct missing image content; or generate a replacement subject.
• Local AI Presenter Anchor uses the latest temporary packaged face landmarks or detected-hand coordinates to position an opted-in text, image, shape, or local video-clip layer. A hand anchor does not require a canned gesture, and Auto/Left/Right selects only the local model's current hand label. Only bounded numeric overlay x/y state is smoothed in memory and it clears on disable, hide/removal, camera changes, reset, or stop. It does not upload or persist frames, landmarks, hand coordinates, handedness labels, face bounds, or motion state; identify a person; inspect audio; infer who is speaking; or generate content.
• Local Smart Presenter Director uses only the current normalized pointer coordinate already sent from the explicitly enabled fixed-tab source helper to choose a padded Canvas corner for an opted-in camera layer. It reads no page text, DOM, URL, click history, or captured pixels; only a bounded numeric camera x/y smoothing state exists in memory and clears on disable, hide/removal, source change, reset, stop, pointer leave, or idle release. It does not upload or persist pointer coordinates or runtime layout state.
• Local AI Camera Denoise Studio performs deterministic edge-aware luma/chroma cleanup at a 320 px maximum long edge. It keeps only one overwritten downscaled derived frame per active camera layer for motion-safe temporal cleanup and clears it on disable, hide/removal, reset, or stop. It does not upload or persist pixels or diagnostics, identify a person, or generate replacement camera content.
• Local Camera Flicker Guard samples only temporary global luminance at a 96 px maximum long edge to damp short LED, fluorescent, or screen-light exposure swings. Its analysis and full-resolution brightness scratch clear after every pass, while bounded numeric baseline/correction state clears on disable, source changes, hide/removal, reset, or stop. It does not upload or persist pixels, samples, or diagnostics; identify a person; inspect audio; determine mains frequency; recover HDR detail; or generate image content.
• Local AI Clip Flicker Guard applies that same capped 96 px luminance analysis and local brightness correction only to an opted-in muted local clip. Its analysis and brightness scratch clear after every pass; bounded baseline/correction state clears on disable, source replacement, hide/removal, reset, or stop. It does not upload or persist clip frames, samples, diagnostics, identities, or source audio; identify a person; inspect audio; determine mains frequency; recover HDR detail; or generate image content.
• Local AI Adaptive Grade measures only a temporary 192 px global luma histogram for an opted-in camera or muted local clip, then applies bounded brightness, contrast, and saturation filters to its original pixels. Its analysis and full-resolution filter scratch clear after every pass; four numeric response values clear on disable, source replacement, hide/removal, reset, or stop. It does not upload or persist frames, luma data, a reference, identities, or source audio; match a reference, apply a LUT, inspect audio, or generate color detail.
• Local AI White Balance measures only a temporary 192 px global RGB sample for an opted-in camera or muted local clip, then applies bounded red, green, and blue gains to its current pixels. Its sample and full-frame pixel buffer clear after every pass; four numeric response values clear on disable, source replacement, hide/removal, reset, or stop. It does not upload or persist frames, RGB data, a reference, identities, or source audio; match a reference, inspect audio, or generate color detail.
• Local AI Color Match keeps only a creator-selected PNG/JPEG/WebP still reference up to 2 MB in local scene settings. For an opted-in camera or muted local clip, it samples temporary 192 px source/reference color statistics and applies bounded RGB/contrast adjustment to current pixels. Sample and full-frame pixel buffers clear after every pass; numeric response state clears on disable, source replacement, hide/removal, reset, or stop. It does not upload footage or the reference; identify a person; inspect audio; apply a LUT; match a remote source; or generate color detail.
• Optional cursor click sounds are synthesized locally and connected only to the recording mix, not directly to your speakers.
• Local trim export routes decoded audio only into the new WebM stream, not to your speakers.
• Local loudness match calculates a temporary average-RMS estimate from the selected recording and applies bounded gain only to a new local export. It does not upload audio, play it through speakers, alter the original, or claim LUFS/true-peak metering.
• AI Smart Slim maps camera pixels through a temporary landmark-guided mesh in Chrome; frames and geometry are not uploaded or retained as biometric templates.
• AI Auto Tone uses a temporary 48 × 48 detected-face crop and bounded color measurements; the crop, measurements, and camera frames are not uploaded, persisted, or used to identify a person.
• Local AI Studio Lights uses temporary packaged-model face bounds to draw bounded procedural lighting locally. It does not generate or replace a face, infer identity, upload pixels or landmarks, or persist face geometry.
• AI Makeup Studio draws procedural color from temporary cheek, lip, and eyelid regions; camera frames, regions, and settings are not uploaded or used to identify a person.
• Aurora, Bokeh Lights, Neon Grid, and Studio Stage are drawn locally from settings and time; they do not fetch stock media or upload camera frames.
• AI Cutout Edge Lab refines temporary person-confidence values and keeps at most one previous 512 px model mask per active camera or opted-in local video-clip layer for temporal stability. Matte Grow / Shrink offsets only the current confidence boundary and creates no added history. It never stores source pixels, uploads or persists a mask, identifies a person, or retains biometric data.
• AI Edge Color Studio uses stateless 320 px scratch copies of the current frame, person matte, and replacement background to detect and suppress edge spill or add light wrap. No pixel, sample, estimate, or correction is uploaded, persisted, added to history, used for identity, or retained as biometric data; only selected scene controls can be saved locally.
• Local AI Scene Match compares temporary 192 px person-interior and replacement-background samples, then caches only bounded numeric exposure, saturation, tint, confidence, and sample-count adjustments. Scratch pixels are overwritten; neither pixels nor derived runtime values are uploaded, persisted, used for identity, or retained as biometric data. Only selected scene controls can be saved locally.
• Local AI Subject FX uses the temporary person mask locally. Motion Echo keeps at most four downscaled Canvas copies for no more than 850 ms, clears them when no longer needed, and never uploads or persists frames or masks.
• Smart Highlights does not upload captions, markers, scores, ranges, or media; selected clips use the same local speaker-isolated export path.
• Local Smart Cut uses temporary local audio-level windows only to suggest silence ranges and draw a bounded 180-bar level map. It does not decode audio a second time, play it to speakers, upload/persist levels, or alter the source recording.
• Instant Replay Export reads only creator-enabled, editable Instant Retake rows after a recording finishes and locally re-encodes each requested clip through the existing speaker-isolated export path. It does not export live media, upload/persist replay data, alter the original, or play decoded audio.
• Local Pause Shortener reuses only the current temporary Smart Cut ranges and analysis-time padding to create gentler editable middle-of-pause suggestions. It never decodes or plays audio again, generates room tone, uploads/persists level data, or alters the original recording.
• Local Rhythm Markers reuse those same temporary level windows to propose spaced, prominent audio-onset timestamps. They do not identify music, infer a beat grid, decode again, play audio, upload/persist analysis values, or alter media; only markers you explicitly add are saved locally.
• Local Retake Finder compares only nearby final local caption starts and cue times to suggest reviewable cuts. It does not inspect audio, identify a speaker, upload captions or ranges, or remove media automatically.
• Local Caption Search & Find compares a typed query only with current local caption rows. It does not upload or persist queries/matches and only seeks the local preview when you select a result.
• Local Caption Cut adds only a cue you explicitly choose to the existing local reviewable cut list. It does not upload or persist caption text/ranges, change the source recording, or create an export until you choose one.
• Local Caption Punch-In Planner compares final local caption timing with already-saved local insight keywords to propose editable Focus Track zoom windows. It does not inspect frames, identify people/speakers/emotions, assess engagement, decode/play audio, upload media, or alter the original recording.
• Local Chapter Cards reuse only the exact time and label of existing creator-approved local markers during a requested export. They do not generate chapter titles, inspect media/audio, upload or persist data, or alter the original recording.
• Local Visual Change Finder samples bounded local video frames only to compare coarse color/brightness changes. You may explicitly add selected candidates as markers or download contiguous local scene clips; it does not identify people or understand content, upload/persist frames or signatures, add a marker/cut automatically, or alter the source recording.
• You may explicitly request up to 12 uncached Visual Change Finder previews at a time. Library serially draws the already-local video preview into one temporary Canvas and keeps only WebP object URLs in page memory; it does not decode audio, upload/persist thumbnails or pixels, and releases previews when analysis or the selected recording changes.
• Local PNG frame export copies only the current local composite preview frame into a temporary Canvas, then downloads it as PNG. It does not upload pixels, inspect audio, or alter the recording.
• Local Storyboard Export serially draws at most 12 local composite preview frames into one temporary Canvas, then downloads a PNG contact sheet. It does not decode audio, upload or retain frames/storyboards, or alter the recording.
• Local Thumbnail Studio ranks caption-timed local moments without analyzing people, scenes, or expressions. You choose the final frame, headline, accent, and PNG format; it uses one temporary Canvas, restores the preview playhead, never decodes audio or uploads/persists pixels or text, and does not alter the recording.
• Local WAV extraction decodes only your selected local recording into the current trim range and downloads 16-bit PCM. Decoded audio is never played, uploaded, or persisted; source recordings over five minutes and WAV output over 128 MB are refused to protect browser memory.
• Failed recording shortcuts store only a bounded local command name, error message, and timestamp so Popup can explain a temporary `!` badge. The value is never uploaded, is ignored after 15 minutes, clears after a successful start, and disappears when Chrome ends.
• Post-production caption/overlay burn-in draws final cue text and optional approved-marker Chapter Cards only into local export canvases and never sends text, frames, or style settings to a server.
• It does not use remotely hosted executable code.
• After you explicitly start Local Whisper or Local AI voiceover, Transformers.js may download the selected model's configuration, tokenizer, and weight data from Hugging Face and cache it in the browser. That network request reveals ordinary connection metadata and the selected public model, but it does not contain recording audio, caption text, or narration text. The ONNX Runtime executable is packaged with the extension rather than loaded from a CDN.
• Vision models are packaged with the extension. Chrome downloads a speech language pack only after you explicitly request it, and caption recognition is forced to local-only mode.
• Subscription status for Pro features is checked with ExtensionPay (extensionpay.com), which handles payment through Stripe. That request carries a random per-install identifier and your paid/unpaid state, and it reveals ordinary connection metadata. It never carries recordings, camera or microphone data, captions, narration text, scripts, or any other capture content. If you never open a Pro feature and never subscribe, the extension still checks this status to decide what to unlock; it sends nothing else. Payment card details are entered only on Stripe's own hosted page and are never seen by this extension or its developer.
• It does not attempt to bypass DRM, HDCP, or website capture controls.

Chrome Web Store Limited Use

Simple Web Recorder's use of information received from Chrome APIs complies with the Chrome Web Store User Data Policy, including the Limited Use requirements. It uses captured website content, the chosen tab's title and URL, and enabled pointer/click activity only to provide the recording, screenshot, editing, recovery, and local OBS-bridge features you explicitly request.

This information is not sold, used for advertising or credit decisions, or transferred for an unrelated purpose. The developer does not receive or allow humans to read your recordings or captured browsing data. The only external requests are the model downloads and payment-status paths described above; captured content is not included in those requests.