# VidProc

> Free video, audio, image and subtitle tools, with local browser previews and a server lane for long or large jobs. Most browser tools decode and re-encode locally with WebCodecs; server-backed exports require an account.

## How it works

- Browser tools process files on the visitor's own device. Nothing is uploaded, nothing is stored, and they keep working offline once loaded.
- The image-sequence slideshow keeps preview and ordering local, then uploads the selected inputs to the authenticated server worker for final rendering and tiered duration limits.
- Where only the container changes, or a cut lands on a keyframe, packets are copied verbatim — lossless and near-instant. Where the picture itself changes, frames are re-encoded. Each tool states which it does.
- A separate server pipeline handles work a browser tab cannot: very long or large files, HDR tone mapping, stabilisation and HLS packaging. That is what the paid plans cover.

- Available in 8 languages: en, zh-CN, ja, es, pt, ko, de, ar.

## Video tools

Convert, compress, trim, reframe, watermark, redact and edit video.

- [hook-checker](https://vidproc.com/en/tools/hook-checker): Score a short video's opening hook 0-100 from its first five seconds: face appearance timing (BlazeFace), speech onset (Silero VAD), visual motion and shot changes (frame differencing), audio energy envelope, and opening-line text patterns when a transcript is available. Findings name concrete fixes (cold open, late face, static visuals). Everything runs in the browser; only an optional 10-second transcription touches the network.
- [watermark-detector](https://vidproc.com/en/tools/watermark-detector): Sample a local video across a chosen interval and detect persistent spatial watermark candidates in a browser worker. Review actual sample evidence, edit and explicitly select up to eight boxes, then pass the local video and validated boxes to the removal editor. Detection is free and uploads nothing; removal retains existing server plan and compute limits. Static backgrounds and moving watermarks are not reliably detected.
- [video-to-gif](https://vidproc.com/en/tools/video-to-gif): Export a clip as an animated GIF or APNG, with region, frame rate and size control.
- [video-preview](https://vidproc.com/en/tools/video-preview): Select or automatically distribute up to ten 2–5 second ranges and combine them into one MP4 preview.
- [video-compressor](https://vidproc.com/en/tools/video-compressor): Reduce file size with a target bitrate or quality, showing the estimated result before encoding.
- [video-trimmer](https://vidproc.com/en/tools/video-trimmer): Cut a clip to a start and end point, on keyframes where possible so no re-encode is needed.
- [video-converter](https://vidproc.com/en/tools/video-converter): Change container and codec. Copies packets verbatim when possible, so it is lossless and near-instant.
- [multicam-sync-editor](https://vidproc.com/en/tools/multicam-sync-editor): Align two to ten camera angles and an optional recorder track by audio envelopes, timecode metadata or manual offsets, preview the sync lanes locally, then render a primary, alternating, picture-in-picture or split-screen MP4 on the server.
- [video-scopes](https://vidproc.com/en/tools/video-scopes): Monitor any video with professional scopes: luma waveform, vectorscope (with skin-tone line) and RGB histogram, updating live during playback or frame by frame on seek, plus luma min/avg/max and black/white clipping readouts. Pure inspection tool, fully in-browser, no upload.
- [pre-publish-checkup](https://vidproc.com/en/tools/pre-publish-checkup): One-click health report for a finished video: pacing curve (cut density per minute), hook score, long-silence and static-stretch hotspots, loudness consistency and subtitle coverage (optional SRT/VTT/ASS), each finding with actionable fixes and click-to-seek timestamps. Pure inspection tool, fully in-browser, no upload.
- [encode-lab](https://vidproc.com/en/tools/encode-lab): H.264 vs VP9 vs AV1 encode lab: encode a short sample of your own video with each codec at the same bitrate, compare file size, SSIM and PSNR side by side, sweep a CRF-style bitrate ladder to find the sweet spot, then hand the winning settings to the video compressor. Free plan benchmarks files up to 10 MB, paid plans unlock any size. Fully in-browser, no upload.
- [transition-studio](https://vidproc.com/en/tools/transition-studio): Add transitions between video clips: cross dissolve, fade through black or white, wipes and slides in four directions, and blur dissolve, each with an adjustable duration (0.1–2s) per junction. Neighbouring clips overlap by the transition, and the audio crossfades in sync (equal-power) over the same window; mixed resolutions and frame rates are normalized to one 720p/1080p MP4, rendered frame by frame in the browser. Up to 8 clips, 15 minutes. Fully in-browser, no upload.
- [seamless-loop](https://vidproc.com/en/tools/seamless-loop): Turn a clip into a seamless loop: auto-detect the loop points whose head and tail frames match best (or trim manually), then build a loop unit whose end crossfades back into its start — or a ping-pong unit that plays forward then backward — and optionally repeat the unit in-file (×1–×10) for players that do not loop. Audio crossfades in sync (crossfade mode) or plays forward-then-reversed (ping-pong; very long segments export silent). Normalized to one 720p/1080p MP4, rendered frame by frame in the browser. Fully in-browser, no upload.
- [cinemagraph](https://vidproc.com/en/tools/cinemagraph): Turn a video moment into a living photo: pick one still frame as the base, brush-paint the region that should move (waterfall, hair, neon), and the painted region plays a seamless loop — crossfade or ping-pong, reusing the seamless-loop phase math — while the rest stays frozen on the still frame. Loop segments are decoded to frames in the browser (max 10 s at 12 fps, capped at 640 px wide), composited per pixel under a feathered brush mask, and exported as a looping GIF (one cycle) or video (3 cycles). Fully in-browser, no upload.
- [beauty-filter](https://vidproc.com/en/tools/beauty-filter): Smooth, brighten and lightly whiten only the face in a talking-head or selfie video: MediaPipe BlazeFace boxes (the same detector auto-redact uses) are tracked live at ~4 fps, smoothed across frames, expanded into a feathered skin mask, and a blurred/brightened copy of each frame is composited back only inside that mask. Strength slider 0–100 (0 is an exact no-op). Preview and export share the same canvas pipeline at up to 1920 px wide; export records the canvas plus the original audio via the video element's media graph, so audio stays in sync. Fully in-browser, no upload.
- [video-reframe](https://vidproc.com/en/tools/video-reframe): Re-crop to a different aspect ratio — 9:16, 1:1, 4:5 — with the subject kept in frame.
- [video-metadata](https://vidproc.com/en/tools/video-metadata): Read container, codec, resolution, frame rate, bitrate and embedded tags; strip tags on export.
- [blur-video](https://vidproc.com/en/tools/blur-video): Blur, mosaic or black out regions. Boxes can follow a moving subject.
- [video-denoiser](https://vidproc.com/en/tools/video-denoiser): Remove noise from night, low-light and old videos: pick light (in-browser canvas filter, preview is exactly what gets exported), medium (server hqdn3d) or strong (server nlmeans). A draggable wipe compares before and after on the same frame, with play, scrub and frame-step transport to check that detail survives. Server levels need sign-in; free accounts are limited to 60-second clips.
- [blur-pad](https://vidproc.com/en/tools/blur-pad): Turn a horizontal video vertical without cropping: the full frame stays centered (contain) while an enlarged, blurred and slightly dimmed copy of the same frame fills the background (cover) at 9:16, 4:5 or 1:1. Output dimensions follow the reframe size rule — no upscale past the source's pixel budget, even dimensions, 1920 px long-side cap — and background blur strength 0–100 maps to canvas blur pixels. Preview and export share the same canvas pipeline; export records the canvas plus the original audio via the video element's media graph. Fully in-browser, no upload.
- [visual-text-search](https://vidproc.com/en/tools/visual-text-search): Search the words printed on screen in a course, meeting or pitch video: frames are sampled at a chosen interval (1/2/5/10 s, stretched past a 300-frame cap) and read in-browser with Tesseract.js (English, Simplified/Traditional Chinese, Japanese or Korean), building a text-to-time index. Consecutive matching samples collapse into one [start,end] occurrence so a persistent slide or caption is a single hit; search matches spaceless text so CJK queries need no tokenizer. Clicking a hit seeks the preview video to the range start, and results export as a CSV timecode list. Frames are extracted in 24-frame batches so only one batch of RGBA is retained, and thumbnails are kept only for frames with text. Fully in-browser, no upload.
- [smart-cover-picker](https://vidproc.com/en/tools/smart-cover-picker): Pick the best cover frame of a video: frames are sampled at a chosen interval (1/2/5/10 s, stretched past a 300-frame cap) and scored in-browser for sharpness, faces (MediaPipe BlazeFace, the same detector auto-redact uses), brightness and rule-of-thirds composition, weighted .35/.30/.20/.15. Ranked covers keep a 2 s minimum gap, then fill back up by score so short clips still yield five candidates; only the top five are re-extracted at up to 1280 px for thumbnails. The winner can be set as the preview video's poster or exported as a JPG. Fully in-browser, no upload.
- [auto-redact](https://vidproc.com/en/tools/auto-redact): Detect faces on-device to seed regions, then track them across the clip so the blur follows the motion.
- [add-watermark](https://vidproc.com/en/tools/add-watermark): Layer text, image and animated-GIF watermarks, marquee text and randomly placed marks, each with its own time window.
- [compare-video](https://vidproc.com/en/tools/compare-video): Play two files side by side or under a wipe, to judge a compression setting.
- [remove-silence](https://vidproc.com/en/tools/remove-silence): Find pauses in speech and cut them out, with the cut drawn on the waveform first.
- [video-speed](https://vidproc.com/en/tools/video-speed): Change playback speed, with pitch-corrected or pitch-shifted audio.
- [video-highlights](https://vidproc.com/en/tools/video-highlights): Find the loudest and most speech-dense moments and export them as short clips.
- [highlights-ai](https://vidproc.com/en/tools/highlights-ai): Find shareable clips by meaning, not loudness: transcribe the video, ask an AI which spans are worth posting, then export each suggestion as its own Short with a title and optional SRT. Free tier transcribes the first 20 seconds.
- [server-processing](https://vidproc.com/en/tools/server-processing): Queue long or large files for server-side processing: HDR tone mapping, stabilisation, deinterlacing, HLS ladders.
- [batch-convert](https://vidproc.com/en/tools/batch-convert): Apply one setting to up to 50 files at once and download the results as a ZIP. Files are processed one at a time locally; a failure is reported per file rather than ending the run.
- [beat-cut](https://vidproc.com/en/tools/beat-cut): Detect musical beats and turn them into cut points, so edits land on the rhythm.
- [zoom-pan](https://vidproc.com/en/tools/zoom-pan): Animate a slow zoom or pan between two framings (Ken Burns).
- [chroma-key](https://vidproc.com/en/tools/chroma-key): Remove a green or blue screen by clicking its colour; handles uneven lighting and green spill.
- [perspective-fix](https://vidproc.com/en/tools/perspective-fix): Straighten footage shot at an angle by dragging four corners onto a whiteboard or document.
- [freeze-frame](https://vidproc.com/en/tools/freeze-frame): Hold a chosen moment on screen, keeping audio in sync afterwards.
- [picture-in-picture](https://vidproc.com/en/tools/picture-in-picture): Composite two clips: corner overlay, side by side or top and bottom.
- [video-join](https://vidproc.com/en/tools/video-join): Merge multiple videos into one MP4 entirely in the browser — mixed resolutions, frame rates and missing audio tracks are normalized: orientation by majority vote, 720p/1080p output box, letterbox or crop fit, per-clip frame-accurate durations, and silent clips filled with generated silence so later audio stays in sync. Scrub the filmstrip to preview the merged result before exporting. No upload, no watermark.
- [video-codec-info-extractor](https://vidproc.com/en/tools/video-codec-info-extractor): Inspect any video locally — codec profile and level decoded from the avc1 string, coded vs display size, rotation, measured frame rate, GOP keyframe distances, a per-packet bitrate chart with I-frame spikes, thumbnails and tags, exportable as a MediaInfo-style text report. Demux-only, no upload.
- [shorts-factory](https://vidproc.com/en/tools/shorts-factory): Split one long video into a batch of platform-ready Shorts: detects highlights by audio energy and by transcript meaning, then exports every clip with titles, aspect ratio, burned-in subtitles, covers and naming rules as one ZIP.
- [brand-kit](https://vidproc.com/en/tools/brand-kit): Put a brand on every video: upload a logo, pick a brand color, and automatically apply an intro sting, a persistent corner bug and an outro end card with a CTA. Save the kit to reuse it across videos, preview with the real compositor and export with everything stamped in.
- [m3u8-downloader](https://vidproc.com/en/tools/m3u8-downloader): Download an HLS (.m3u8) stream as one MP4 file, entirely in the browser — nothing is uploaded. Pick a quality from a master playlist, fetch the segments, decrypt AES-128 when the key is served openly, and remux to MP4. Mid-roll ads are detected (SCTE-35 cues, ad-CDN host divergence, or discontinuity alternation) and removed automatically.
- [review-studio](https://vidproc.com/en/tools/review-studio): Review videos with your team and clients: create a project, upload review versions, and share an unguessable link that viewers can comment on without an account. Comments carry a playback timestamp (and optional on-frame box), can be flagged as to-dos for the editor, and each version keeps its own thread — one project library, long-lived files, Range-streamed playback.
- [compliance-check](https://vidproc.com/en/tools/compliance-check): Run a pre-publish compliance check for TikTok, Reels, Shorts, YouTube, Bilibili or podcast specs: container and codec acceptance, aspect ratio, resolution, frame rate, duration limits, bitrate headroom, LUFS loudness against the platform target, peak ceiling, and — with an SRT/VTT/ASS loaded — subtitle timing, reading speed, line length and cue duration. Everything is measured locally and returned as a pass/warn/fail report you can download as Markdown or JSON.
- [workflow](https://vidproc.com/en/tools/workflow): Subscriber-only pipeline on a free-form drag canvas: wire trim, redact, watermark, transcode, audio and artifact (GIF, frames, screenshot, thumbnail, preview) nodes into any acyclic graph — one source can fan out to several outputs on independent paths — and every chain still compiles into a single encode. Redact regions are drawn directly on the video frame (same visual editor as the standalone tool), tuned node parameters save as reusable per-kind presets, and a natural-language AI bar proposes nodes/parameters from your description (every proposal re-validated by the editor's rules; only media metadata is sent, never the file). Files never leave the device.
- [on-screen-text-translator](https://vidproc.com/en/tools/on-screen-text-translator): On-screen text translator: sample a video at a chosen interval (1/2/5 s, stretched past a 300-frame cap) and read the burned-in words in-browser with Tesseract.js (English, Simplified/Traditional Chinese, Japanese, Korean, or English + Simplified Chinese), grouping words into lines and collapsing the same text in nearly the same box across samples into one tracked item — one missed sample does not split it, and a moving box keeps a time-stamped track that interpolates between samples. Each item is reviewable: edit the original or the translation, adjust its time range, delete it, or drag or resize its box on the preview via the shared BoxOverlay (the new position is saved at the playhead time). Translation runs through the existing subtitle-translation endpoint, so only the text lines leave the browser. Export burns the translated bars — translucent dark bars with wrapped, fitted text — over the frame with the shared per-frame render pipeline and keeps the original audio; clips over 15 minutes can be scanned and reviewed but not exported in the browser. No server render route in this version.
- [telegram-link-downloader](https://vidproc.com/en/tools/telegram-link-downloader): Telegram link downloader: link your own Telegram account (phone number, login code, two-step verification password when enabled) and paste one t.me message link — private (t.me/c/...), comment, or public (t.me/channel/...) — to download its videos. The download runs on the server through your account: albums download as a group, files are named from the message text plus the original video name with _01, _02 suffixes, and you get HMAC-signed download links valid for 12 hours, after which the files are deleted. One download runs at a time per account and extra jobs wait in a FIFO queue; downloads stop after 60 minutes. Free accounts get one trial download per account, Plus 20 jobs a day, Pro 100. Your Telegram session is AES-256-GCM encrypted on the server and never sent to the browser; it can be revoked here or inside Telegram at any time.

## Audio tools

Extract, clean, level, trim and mix audio.

- [audio-cleaner](https://vidproc.com/en/tools/audio-cleaner): Reduce noise, remove hum, gate breaths and level speech, previewed against the original.
- [podcast-studio](https://vidproc.com/en/tools/podcast-studio): Master a single-track podcast on the server: voice cleanup, EQ, compression, silence cutting for audio-only inputs, LUFS alignment, and show notes.
- [extract-audio](https://vidproc.com/en/tools/extract-audio): Pull the audio track out of a video as MP3, M4A, WAV, FLAC or OGG.
- [video-volume](https://vidproc.com/en/tools/video-volume): Measure peak and RMS level, then normalise or apply a gain, with fade in and out.
- [audio-trimmer](https://vidproc.com/en/tools/audio-trimmer): Trim audio on an interactive waveform.
- [segment-gain](https://vidproc.com/en/tools/segment-gain): Segment gain studio: pick ranges on the waveform, set a linear gain per range (0% silences, 100% unchanged, up to 200%), and export. Segment edges crossfade over 0-0.5 s (default 0.1 s); gaps shorter than the crossfade blend directly into the next segment instead of dipping to 1, and overlapping segments resolve in time order with the later segment winning. Preview plays the exact envelope the export renders (WebAudio gain automation sampled every 10 ms, the same schedule applyVolumeEnvelope uses). Local export renders in-browser and never uploads; long files can go to the server route instead, which reuses the Plus/Pro subscription gate (free users get 402 with an upgrade hint) and applies the identical curve as an ffmpeg per-frame volume expression, returning MP3 with no files stored.
- [de-breath](https://vidproc.com/en/tools/de-breath): De-breath studio: detect audible inhales between phrases and mute or soften each one, without cutting anything so pauses keep their length. Detection runs in-browser: Silero VAD speech ranges carve the clip into gaps (an energy mask is the fallback when the model cannot load), and inside each gap a breath is a run of frames that is loud for that gap, noise-like (high spectral flatness) and concentrated in the 300-6000 Hz band — never a volume gate alone, and word tails are excluded by padding around speech. Each breath is reviewable: audition it in context, set attenuation 0-100% (100% mutes), restore it, or delete it; sensitivity re-detects from the cached analysis. Preview plays the exact envelope the export renders. Local export renders in-browser and never uploads; long files can go to the server route instead, which reuses the Plus/Pro subscription gate (free users get 402 with an upgrade hint) and applies the identical curve as an ffmpeg per-frame volume expression, returning MP3 with no files stored. Detection always runs in the browser; the server only renders the reviewed segments.
- [audio-join](https://vidproc.com/en/tools/audio-join): Merge several audio files into one, with an optional crossfade at each join capped so it cannot swallow a short clip.
- [audio-speed](https://vidproc.com/en/tools/audio-speed): Speed up or slow down audio. Pitch can be preserved (WSOLA time-stretch) or allowed to shift like a tape machine.
- [pitch-shift](https://vidproc.com/en/tools/pitch-shift): Shift audio pitch by semitones (±12, one octave) without changing tempo or duration: resample by the pitch factor, then WSOLA time-stretch back to the exact original length. Fun-voice presets are pitch-only, and the first 10 seconds can be previewed through the real engine before export. Fully in-browser, no upload.
- [audio-channels](https://vidproc.com/en/tools/audio-channels): Stereo to mono, mono to stereo, swap sides, copy one side to both to rescue a dead microphone, or extract mid or side.
- [audio-silence](https://vidproc.com/en/tools/audio-silence): Find and cut the pauses in a podcast or voice recording, shown on the waveform before committing. No length cap beyond an hour.
- [audio-eq](https://vidproc.com/en/tools/audio-eq): Five-band EQ with a visible curve plus a compressor, with presets for voice, podcast and rumble reduction and automatic make-up gain.
- [audio-fade](https://vidproc.com/en/tools/audio-fade): Fade in and out by the numbers, or draw a volume curve directly on the waveform.
- [vocal-remover](https://vidproc.com/en/tools/vocal-remover): Split any song or video's audio into instrumental (karaoke) and vocals (acapella) entirely in the browser — median spectral subtraction with the spectrogram as the visual editor: drag the vocal band edges where you see the voice, audition both stems side by side, export WAV/MP3/FLAC or both in a ZIP. No upload, no model download.
- [audiogram](https://vidproc.com/en/tools/audiogram): Turn audio or a podcast clip into a social-ready vertical (9:16, 1:1 or 16:9) waveform video entirely in the browser — pick the 15–60s window on the timeline, add cover art as a blurred backdrop or rounded card, title text and a scrolling or full waveform with progress, and export a local MP4 with the audio muxed in. No upload, no watermark.
- [duck-music](https://vidproc.com/en/tools/duck-music): Lower a music bed automatically while someone is speaking.
- [loudness](https://vidproc.com/en/tools/loudness): Measure integrated loudness in LUFS (ITU-R BS.1770) and match a platform target.

## Image tools

Pull stills, frame sequences, sprite sheets and animated images out of video; resize and re-encode still images in batches.

- [manga-translator](https://vidproc.com/en/tools/manga-translator): Upload one manga page, draw speech-bubble regions, run local CPU OCR, translate editable text, and export a clean PNG. The MVP is designed for light, mostly flat bubbles.
- [video-to-frames](https://vidproc.com/en/tools/video-to-frames): Export every frame, or every nth frame, as a ZIP of PNG or JPEG stills.
- [image-sequence-video](https://vidproc.com/en/tools/image-sequence-video): Preview and order multiple images or a ZIP in the browser, then render a server-side MP4 slideshow with per-image timing and optional looping audio; paid plans allow timelines beyond the browser's 15-minute local cap.
- [animated-webp-video](https://vidproc.com/en/tools/animated-webp-video): Upload and arrange animated WebPs, play each animation once, automatically select landscape or portrait by source counts, and sequentially encode a complete H264 MP4 on a FFmpeg 9 server. Free users accept ten files per job; subscribers have no file-count cap and consume measured PFU compute allowance. Upload totals and server timeout still apply.
- [batch-image-resizer](https://vidproc.com/en/tools/batch-image-resizer): Resize up to 50 images at once in the browser: by width, height, percentage (10-200%) or preset (avatar 512 px, social post 1080 px, product detail 750 px, social cover 1200 px), with the aspect ratio locked or stretched, and download the resized set as a ZIP. Target sizes are shown on every card before running; files already at the target size in an encodable format are copied, not re-encoded. Output keeps the source format (PNG, JPEG, WebP); AVIF and GIF export as PNG (GIF first frame only). Fully in-browser, no upload.
- [video-screenshot](https://vidproc.com/en/tools/video-screenshot): Grab a single frame at an exact timestamp.
- [cover-pick-server](https://vidproc.com/en/tools/cover-pick-server): Subscriber-only server cover pick: upload up to 10 stills (Plus/Pro gate; free users get 402 with an upgrade hint) and the server scores each with sharp (EXIF auto-rotate, 480 px) for sharpness, subject prominence (Sobel edge mass: central concentration, outline-scale presence gate, coverage), brightness and rule-of-thirds composition, weighted .35/.30/.20/.15, then ranks them best-first. The subject factor replaces the browser picker's face factor because MediaPipe BlazeFace is not available in the Node route runtime — so 2D/3D animation, real people and vertical screenshots share one rule. Images are scored, not stored; no job is queued.
- [gif-to-video](https://vidproc.com/en/tools/gif-to-video): Convert an animated GIF into MP4 or WebM, which is far smaller for the same clip.
- [duplicate-photo-finder](https://vidproc.com/en/tools/duplicate-photo-finder): Find visually duplicate photos in a whole folder without uploading anything — three perceptual hashes (aHash, dHash, pHash over a DCT) group resized, recompressed and color-graded copies of the same shot, each group recommends keeping the highest resolution, and Chrome can delete the discarded originals in place (else export a CSV cleanup manifest). HEIC files the browser cannot decode are skipped with a clear reason. Photos never leave the device.
- [video-sprite-sheet](https://vidproc.com/en/tools/video-sprite-sheet): Build a sprite sheet plus a WebVTT file for video-player scrub thumbnails.

## Subtitle tools

Transcribe, translate, edit and convert subtitles.

- [subtitle-qa](https://vidproc.com/en/tools/subtitle-qa): Check SRT, VTT and ASS subtitle files: timeline overlaps, minimum cue gaps, reading speed (CPS), duration bounds, line length and line count, empty cues, and spacing/CJK punctuation hygiene. Every fixable issue offers a one-click repair plus fix-all; exports back in the source format. Fully in-browser, no upload.
- [subtitle-editor](https://vidproc.com/en/tools/subtitle-editor): Edit subtitle timing and text against the video.
- [subtitle-dubbing](https://vidproc.com/en/tools/subtitle-dubbing): Translate SRT/ASS subtitles, review cue timing, generate a timed AI voice track, and export a dubbed or dual-track video. Free registered accounts are limited to 20 seconds per clip.
- [danmaku-to-subtitle](https://vidproc.com/en/tools/danmaku-to-subtitle): Turn a Bilibili danmaku file (XML or JSON) into an editable subtitle file: comments are sorted, filtered (sensitive words, repeat spam within 30 s, per-second burst caps) and windowed into timed cues — scrolling comments get 8 s, pinned top/bottom comments 4 s — then exported as SRT or VTT plain timed lines, or as ASS with lanes: the danmaku track layout assigns scrolling lines to non-overlapping traversal lanes (\move) and pinned lines to rows (\an8/\an2 \pos) on a 1920×1080 grid. Preview table lets you edit text or delete cues before exporting. Fully in-browser, no upload.
- [subtitle-muxer](https://vidproc.com/en/tools/subtitle-muxer): Mux SRT, VTT or ASS subtitles into a video as switchable tracks, without re-encoding: video and audio packets are copied unchanged with mediabunny (rotation, languages and track names carried over; packets with negative timestamps, i.e. AAC priming frames, are dropped), every subtitle file is normalized to WebVTT first (ASS styling is lost, only text and timing survive), and each track gets an ISO 639-2/T language code, a name and an optional default flag. MP4 output carries WebVTT as wvtt, MKV as S_TEXT/WEBVTT. Subtitle tracks already in the source are not preserved: this mediabunny version cannot read subtitle tracks out of any container, so only video and audio are copied. Verification is byte-level, not a reopen: the output's container bytes are scanned (MP4 box walk, MKV EBML walk) for the subtitle tracks, because mediabunny cannot read back what it just muxed. Note: browsers' built-in players do not render muxed subtitle tracks — VLC/mpv do. Fully in-browser, no upload.
- [subtitle-converter](https://vidproc.com/en/tools/subtitle-converter): Convert between SRT, VTT, ASS and SSA.
- [subtitle-extractor](https://vidproc.com/en/tools/subtitle-extractor): Extract embedded text subtitle tracks from MKV, MP4 and MOV files and export them as SRT or VTT. Lists every subtitle track with language and codec; picture subtitles (PGS, VobSub) are listed but not exported. Fully in-browser, no upload.
- [subtitle-burn](https://vidproc.com/en/tools/subtitle-burn): Burn SRT, VTT or ASS subtitles permanently into a video with caption styling and platform safe areas for TikTok, Reels, Shorts and Bilibili. Optionally export the subtitle file alongside.
- [danmaku-burn](https://vidproc.com/en/tools/danmaku-burn): Burn Bilibili-style danmaku (bullet comments) into video: load an XML or JSON danmaku file, tune scroll speed, density, font size and pinned-comment duration, preview with the same compositor the export uses, and burn scrolling plus top/bottom comments permanently into the video. Fully in-browser, no upload.
- [text-cut](https://vidproc.com/en/tools/text-cut): Edit video by editing text: transcribe speech with AI, strike the words or sentences you don't want, and export the cut video plus matching subtitles. Free tier transcribes the first 20 seconds.
- [auto-subtitles](https://vidproc.com/en/tools/auto-subtitles): Transcribe speech to subtitles, translate them, and burn or export SRT/VTT.
- [ass-editor](https://vidproc.com/en/tools/ass-editor): Edit ASS subtitle files with full inline-effect support: fonts, colours, fades, moves, 	 animations and karaoke, with a live canvas preview and burn-in export through the same renderer.

## File tools

Rename, organize and manage files locally in the browser.

- [pdf-merge](https://vidproc.com/en/tools/pdf-merge): Combine two or more PDFs in the exact visible order with local page inspection and preview. The browser copies original pages without rasterizing them, preserves dimensions, rotation and selectable text, and downloads one merged PDF without uploading source files.
- [pdf-split](https://vidproc.com/en/tools/pdf-split): Extract chosen pages into one PDF or split one document into named PDF parts inside a ZIP. Supports strict ranges plus odd, even and every-N selectors; pages are copied without rasterization and source files stay in the browser.
- [pdf-organize](https://vidproc.com/en/tools/pdf-organize): Reorder, rotate and remove PDF pages with on-demand thumbnails, multi-select actions and up to 50 undo steps. The browser copies original PDF pages rather than screenshots and keeps the source file local.
- [images-to-pdf](https://vidproc.com/en/tools/images-to-pdf): Arrange PNG, JPEG and WebP images and create A4, Letter or original-size PDF pages locally with contain or cover fit, margins and background color.
- [pdf-to-images](https://vidproc.com/en/tools/pdf-to-images): Render selected PDF pages to PNG, JPEG or WebP at an exact DPI. One page downloads directly and multiple pages are packed into an ordered ZIP.
- [pdf-to-word](https://vidproc.com/en/tools/pdf-to-word): Convert a PDF into an editable .docx in the browser: text clusters into paragraphs, large type becomes Word headings, aligned columns become real tables, and embedded images stay in place. Up to 5 PDFs per queue; scanned pages have no text layer and convert empty with a warning to run pdf-ocr first.
- [pdf-to-excel](https://vidproc.com/en/tools/pdf-to-excel): Convert a PDF into a real .xlsx spreadsheet in the browser: aligned columns across three or more rows become spreadsheet tables, cells are typed as numbers, dates or text, and sheets can be one per table, one per page, or a single merged sheet. Up to 5 PDFs per queue; scanned pages have no text layer and convert empty with a warning to run pdf-ocr first.
- [pdf-watermark](https://vidproc.com/en/tools/pdf-watermark): Add multilingual text or PNG, JPEG and WebP image watermarks to selected PDF pages with opacity, rotation, scaling, nine-point placement or deterministic tiling. The browser embeds the mark without rasterizing original pages.
- [pdf-page-numbers](https://vidproc.com/en/tools/pdf-page-numbers): Add multilingual headers, footers and page numbers to selected PDF pages. Safe variables expand the page number, selected-page total, date and filename; text is embedded locally without rasterizing original pages.
- [pdf-compress](https://vidproc.com/en/tools/pdf-compress): Flatten image-heavy or scanned PDFs into controlled-resolution JPEG or WebP-quality page images, then rebuild a compatible PDF locally. The result reports exact before-and-after sizes and warns that searchable text, links, forms and vectors become pixels.
- [pdf-ocr](https://vidproc.com/en/tools/pdf-ocr): Turn scanned PDF pages into searchable documents with local CPU OCR. Existing text pages are skipped by default; selected scan pages retain their original appearance while an invisible multilingual text layer is added.
- [pdf-security](https://vidproc.com/en/tools/pdf-security): Protect a PDF with AES-256 permission flags or unlock one using a password the user already knows. Processing is authenticated and server-side; passwords are kept only in a short-lived restricted temporary file and never stored in job records.
- [pdf-workflow](https://vidproc.com/en/tools/pdf-workflow): Compose compatible PDF and image operations as a typed sequence. Browser stages run locally, while AES-256 security uses an authenticated short-lived server job; templates intentionally exclude files and passwords.
- [bulk-rename](https://vidproc.com/en/tools/bulk-rename): Rename many files locally in your browser with a pipeline of rules (replace, regex, case, add/remove text, numbering, extension) that previews in real time with per-file diffs and conflict detection. Pick a folder and rename in place, or pick files and export a rename manifest JSON — nothing is uploaded.

## Other pages

- [All tools](https://vidproc.com/en/tools)
- [Pricing](https://vidproc.com/en/pricing)
- [Terms](https://vidproc.com/en/legal/terms)
- [Privacy](https://vidproc.com/en/legal/privacy)