Studio guide

Getting started

1. Open the studio and drop one or more recordings (WAV, MP3, OGG, FLAC or AIFF), a whole folder, or a video (MP4, MOV, WebM): Otocut splits its sound, and can export video clips (see Video).
2. Otocut immediately finds every sound separated by silence. Listen to them and fix anything you need.
3. Name them in the Naming tab, set up the Export tab, then click Export (.zip).

No file at hand? The Try with a demo pack link loads a synthetic sound effects recording: footsteps, whooshes, glass, impacts and a wind ambience.

Lost? Ctrl + K opens the command palette: type what you want to do (“export”, “tune”, “map”…) and press Enter. The guided tour, offered the first time you open the studio, can be started again from the palette.

With a folder, the original subfolders are kept in the .zip. Untick Keep the original subfolders in the export settings to flatten everything.

Simple and advanced mode

The Simple · Advanced switch, top right, chooses what the studio shows. Simple mode keeps the essentials: the silence threshold, the sound titles, a sound's volume and pitch, the target engine and the export format. Advanced mode shows every setting (cleanup, variations, EQ, stereo, tuning, UCS, map, recipes, mic and generator…).

• New visitors start in simple mode. The studio remembers your choice.
• Advanced settings stay applied in simple mode: a box in the Export tab lists the ones that change the export.
• The Ctrl + K palette also finds advanced features; picking one switches to advanced mode.
• In simple mode, the first time you open the studio, Otocut asks what your sounds are for (video or podcast, Unity, Godot, Unreal, web game, RPG Maker, sound library) and sets up the export accordingly. With “Video or podcast”, detection cuts at every sentence and the voice is cleaned up and brought to podcast loudness (−16 LUFS). Change your answer from the palette (“Choose what my sounds are for”).

On tablets and phones, the studio shows one panel at a time: Tracks, Sounds or Settings, picked at the bottom of the screen. Simple mode is the most comfortable there.

Detection

Detection looks for passages louder than the silence threshold, separated by at least the minimum silence.

Silence threshold

Anything quieter counts as silence. The Auto button sets it from the recording's background noise. Raise it if the hiss is loud.

Minimum silence

Lower it if two close sounds end up as one; raise it if a sound (footsteps, a rumble) gets cut into pieces.

Minimum sound

Shorter sounds are ignored: clicks, stray noises.

Padding

Silence kept before and after each sound, so the attack and the tail aren't cut.

Presets cover the common cases (short sounds, long sounds, loud background noise) and you can save your own. All recordings runs detection everywhere with the same settings.

Editing

• Create a sound: drag over an empty area of the waveform.
• Adjust: drag the edges of a sound.
• Split at the cursor with S, merge with the next one with M, delete with Del. Right-click offers the same actions.
• Zoom with the mouse wheel, 0 to see everything again. The spectrogram button shows frequencies under the waveform.
• Ctrl + Z undoes, without limit.

Export preview plays each sound with the export settings applied (normalization, fades, mono, sample rate, random variation).

Sound settings

Select a sound then open the Sound tab (or right-click, Sound settings). Its settings add to the export settings and end up in the exported files; a small wand next to its number in the list shows them.

Volume and pitch

From -24 to +12 dB, and from -12 to +12 semitones (+12 raises it by an octave and halves its length).

Tuning

Find the note shows the note the sound plays (A4, C5… with the offset in cents). Tune to snaps it to the nearest note, or to a scale (major, minor, pentatonic) from the chosen root: blips, chimes or piano notes sound in tune together. The pitch you set stays in the scale. Tune the whole track does it for every sound with a clear note; noises (footsteps, impacts) stay as they are.

Exact length

A length in milliseconds: the sound is stretched or tightened without changing its pitch, and the exported file lasts exactly that long (loops too, with their seam kept). With a pitch set, only the note changes. The exact length is shown in color in the list.

Stereo

Pan places the sound left or right; width narrows (0% = mono) or widens the stereo image. For a mono sound, above 100%, pseudo-stereo opens it up without losing anything when played back in mono. The sound then comes out in stereo, unless you export in mono.

Fades

A fade in or fade out for this sound only. Left empty, the Export tab's fades apply.

Equalizer and compression

Bass (120 Hz), mids (1 kHz) and treble (6 kHz), from -12 to +12 dB. Compression brings loud and quiet parts closer: the sound feels louder and punchier, without going past its original peak.

Reversed

The sound played from the end to the start: an incoming whoosh, a sucked-in cymbal.

Test bench

B key: the sound, the sounds of its group and their variations are replayed at random, at a game's pace (80 ms to 1.2 s between sounds). You hear right away whether the repetition stands out. With no variations set, the bench makes up four to let you hear what they would bring.

Similar sounds

Shows in the library the twelve sounds most like this one, across all recordings, with their similarity as a percentage. Otocut compares tone and length, and the kind of sound when AI recognition is available. Also from the right-click menu or each thumbnail's button.

Layer

Mixes the sound with another one, from any recording: an impact and debris, a door and a creak. Set the second sound's offset and each volume, listen, then Create the sound adds it as a new recording.

Keyboard review

To sort a big pack quickly, press R (or the Review button). Each sound plays in turn:

Naming

Numbered title

Every sound gets the title followed by a number: door_01, door_02… The first number and the number of digits are adjustable.

Template

The shape of the names, with the tags {title}, {n} (number) and {track} (source file name). For example SFX_{title}_{n}.

One name per sound

Type each sound's name in the list, or paste a list (one line per sound, in order).

Repeated names

With Number repeated names, three sounds named step become step_01, step_02, step_03.

Find variations

Groups sounds that are alike (same timbre, same length) and gives them a shared name. Renaming one sound renames its whole group.

Suggest names

A sound recognition AI (Google's YAMNet) listens to each sound on your computer and suggests a name: footsteps, door, glass, explosion… Hover a name to see what it heard. Always check its suggestions.

Find and replace

Changes a word in titles and typed names, in the recording or in all of them. Case is ignored; with Regular expression, ^ adds a prefix (^ replaced with SFX_) and _\d+$ removes a trailing number.

UCS naming

The naming convention of professional sound libraries (Universal Category System): CatID_Name_Creator_Source, for example DOORWood_door 01_SIMON_MYPACK. Turn on UCS naming, type your creator ID and your library name (the source, the recording's title by default). Each sound gets its category (CatID) in the sound list; Suggest categories guesses them with the AI, and the default category is used for sounds without one. The built-in list covers the most common categories: you can type any other CatID from the official UCS list.

Variations and loops

Automatic variations

In a game, hearing exactly the same sound on every step or hit gets old fast. In the Export tab, Variations per sound adds slightly higher or lower versions of each sound (Pitch ±, in semitones), a little quieter (Volume): step_v1, step_v2… The game engine picks one at random. Tone ± also makes them slightly brighter or duller, and Length ± slightly longer or shorter, without changing their pitch. Variations stay the same for a given sound from one export to the next.

Seamless loops

For an ambience (wind, rain, crowd, engine), mark the sound as a loop with the ⟳ icon in the list or the L key. On export, the end of the sound is crossfaded into its start, at the point where they match best: the loop runs without clicks or gaps. Loop crossfade sets the length of that crossfade (200 ms by default, longer for very changing ambiences). A loop plays three times in a row in the studio so you can check the seam.

Export

Formats

16, 24 or 32-bit WAV, MP3 or OGG Vorbis, at the original sample rate or at 22.05, 44.1 or 48 kHz, stereo or mono.

Normalize

In dB peak, every sound reaches the same maximum level. In LUFS, every sound has the same perceived loudness (ITU-R BS.1770), never above -1 dB peak: what you want for a consistent pack.

Low cut

Removes low rumble: wind, mic handling, ventilation.

Trim leftover silence

Removes what's left below the threshold at the start and end of each sound.

Hiss reduction

Lowers background noise (mic hiss, ventilation) by 6, 12 or 18 dB. Otocut measures it in the gaps of the recording, then only lets through what clearly rises above it. Without noisy gaps (synthesized pack, digital silence), there's nothing to remove and the sound stays untouched.

Remove clicks

Erases very short crackles (2 ms at most): digital glitches, mouth clicks, vinyl dust. Avoid it on sounds made of crackles, like fire or frying. Listen with Export preview to compare.

Metadata in WAV files

Each WAV carries its name, UCS category, tags, creator and library, in the INFO, BWF (bext) and iXML chunks read by Reaper, Soundminer and most sound library managers.

Fades

Fade-in and fade-out length, and their shape: Linear, Smooth (S-shaped, no harsh attack) or Fast (the sound drops quickly, then fades gently; good for impacts).

Quality check

The button under Export measures every sound with the export settings (length, peak, LUFS) and flags what deserves a listen: clipped in the recording, clipped by too much volume, very quiet, very short, DC offset, cut attack or cut tail. Hover a problem to see how to fix it, click a sound to listen to it, and keep the report as a CSV.

Audio sprite

All sounds in a single file, with its JSON maps for Howler.js and Phaser.

Favorites only

Exports only the sounds marked with a star.

Also export for…

Tick other engine presets: each gets its own folder in the same .zip, with its own settings. Handy to deliver a Unity and a Web pack at once.

Game engines

Pick an Engine preset at the top of the Export tab: it sets the right format and sample rate.

Unity

16-bit WAV at the original sample rate. Drag the folder into Assets with its .meta files. A sound and its variations become an AutocutSound asset that plays a random one, never the previous one:

public AutocutSound door;
door.Play(audioSource);

The AutocutSound.cs script is included; keep a single copy if you import several packs.

Unreal Engine

16-bit WAV at 48 kHz. Import the files into the Content Browser.

Godot

16-bit WAV at 44.1 kHz. Drag the folder into the project. A sound and its variations become an AudioStreamRandomizer (a .tres file): put it in an AudioStreamPlayer's Stream, each play() picks a random variation. For a loop, turn on Loop Mode in the WAV import options.

RPG Maker MV / MZ

OGG at 44.1 kHz. Copy the files into audio/se.

Web (Howler.js, Phaser)

An MP3 sprite at -16 LUFS and its maps. With Howler.js:

const sfx = new Howl(await (await fetch("pack.howler.json")).json());
sfx.play("door_01");

With Phaser 3: this.load.audioSprite("sfx", "pack.phaser.json", ["pack.mp3"]). Loops are already marked in both maps.

FMOD Studio

24-bit WAV at 48 kHz, one folder per sound and its variations. Drag the folder into the Assets tab. For a sound with variations, select all the files in its folder and drag them together onto an event track: FMOD creates a Multi Instrument that plays a random one. The .zip includes instructions.

Wwise

24-bit WAV at 48 kHz and a wwise_import.txt file. In Wwise, open Project > Import Tab Delimited… and pick that file: each sound becomes a Sound SFX, and a sound with its variations lands in a Random Container. In Wwise 2024 or later, the Actor-Mixer hierarchy is called Containers: replace \Actor-Mixer Hierarchy\ with \Containers\ in the file before importing.

Video

Drop a video (MP4 or MOV) like an audio recording: Otocut finds the silent moments the same way, and the picture shows above the waveform. It follows playback; the film button, next to the zoom, hides it.

In the Export tab, the Video box picks what the .zip holds, on top of the sounds:

• One video clip per sound: one MP4 per sound, picture and sound, named like the sound. Volume, fades, EQ and cleanup apply; pitch, exact length, reverse and loops don't, so the sound stays in sync with the picture.
• The video without the silences: all the sounds end to end in a single video, and the same as audio only (handy for a podcast or an interview).
• Editing files: the cuts, without the silences, to open in Premiere Pro and DaVinci Resolve (.xml), Final Cut Pro (.fcpxml) or any editor (.edl). They point at your original video: nothing is re-encoded, and you can extend each cut. README_EDITING.txt explains the import in each editor.
• One animated GIF per sound: a small animation (8 seconds at most), to share on Discord or a website.
• Subtitles (.srt): speech is recognized on your computer (Whisper), for the original video, the video without silences and each clip. Pick the spoken language and the accuracy: the model (40 or 80 MB) downloads only the first time you tick the box and export, then stays in your browser.
• Subtitles burned into the picture: the words show in the clips, the video without silences and the GIFs, a few at a time, the spoken word in yellow, like on Shorts and TikTok. Each word's timing is estimated from the recognized sentence.
• Transcript and chapters (.txt): all the recognized text with timings, and chapters ready to paste into a YouTube video description. They are cut at the longest silences between sentences, about one every 90 seconds, and titled with their first words: proofread and rename them if needed. Timings follow the video without silences when it is exported, otherwise the original recording.
• The sounds as audio files: untick it to get only the video.
• Picture format: original, vertical 9:16 (Shorts, TikTok, Reels) or square. The Framing slider moves the kept picture left or right.
• Silence kept between sounds: instead of cutting everything, keep for example 400 ms of silence between two passages, for a more natural edit (video without silences and editing files).

Audio podcasts: with the “Video or podcast” use (or the “Video and podcast options” palette command), the box also shows for an audio file, as “Podcast and speech”: the audio without silences, subtitles and transcript work the same way. Speech is recognized once per recording and kept with your session: splitting or deleting sounds doesn't redo it, changing the language or accuracy does.

Audiograms: for an audio file, “One audiogram per sound” turns each sound into a video to share: bars that follow the voice, the recording's title, a progress bar and, if you tick them, burned-in subtitles. Pick the format (landscape, vertical 9:16 or square) and, if you like, a background image such as the podcast cover. Keep only the parts you want to share (or tick “Favorites only”): each sound becomes a video.

Cutting the “uh”s: in the Detection tab, “Find hesitations” looks for the “uh”, “um” and “mmm” in the recording's sounds: a vowel held without changing pitch or tone. Listen to each one found, untick the ones to keep, then cut: a sound cut in the middle becomes two sounds, and the video without silences joins them back. Ctrl+Z undoes it.

Edit by text: the “Text” view, next to the sound list, shows the transcript sentence by sentence. Click a sentence to hear it, ✕ to remove it from the sounds (it stays listed, struck through), ↶ to bring it back. The first time, “Transcribe” runs speech recognition on your computer.

Interviews with several mics: when you add several recordings of the same length (one mic per voice, recorded together), Otocut offers to “Mix the interview”. It turns them into a single track: every voice is brought to the same volume, and the mic of whoever is silent is turned down to cut the hiss and the echo of the other voices. If one of the recordings is a video, the interview keeps its picture for the clips and the video without silences. The original recordings lose their sounds (Ctrl+Z brings them back). The mics must have started recording at the same time.

Long episodes: with Chrome or Edge, tick “Write to a folder instead of a .zip” in the Export tab. Files are written one by one into the folder you choose, and videos frame by frame: nothing stays in memory, even for an hour of video. Speech recognition uses several processor cores, and the graphics card when the browser allows it (for more than 2 minutes of speech; its model is a bigger download the first time).

The sounds of a video get a thumbnail in the list and in the library. In the Naming tab, Name from speech gives each sound its first spoken words (same model as the subtitles).

Videos are written by your browser, with nothing uploaded, as MP4 (H.264 when it can). You need an up-to-date browser: Chrome and Edge work best. Otherwise the audio and the editing files still export, and the box tells you. WebM and MKV files only give their sound.

Studio

The Studio edits a short video from your passages: a Short, a podcast clip, a trailer. Open it with the Studio button at the top of the screen, or Edit in the Studio in the Video box of the Export tab. The first time, the timeline fills up with the detected passages of your videos, in order. The Otocut button, top left, takes you back to the splitting screen.

• Media (on the left): Import adds videos, music or images without going back to the splitting screen (you can also drop them on the studio). A video with no sound, such as a screen recording, is added silent. The bin next to a media item removes it, with the shots that use it; a video or a sound also leaves the splitting screen. For each video, Passages adds the detected sounds, Whole the whole video, Cutaway puts it full frame on a video track above, Overlay in a corner of the picture. An audio track is added as Music (looping, ducked under the voice) or as a Sound (at the playhead, on a free audio track). The Image button adds a photo or a logo, as a Shot with a slow zoom or as an Overlay. The Title and Blur zone buttons add an item at the playhead.
• Timeline: drag a shot to change the order, pull its edges to shorten or lengthen it. Titles and effects sit at the very top. Click the ruler to move the playhead, pull the top edge of the timeline to make it taller.
• Tracks: V1 holds the shots, the video tracks above (V2, V3…) the overlays and cutaways (the highest one is in front), the audio tracks (A1, A2…) music, voices and sounds. + Video and + Sound add a track. Drag a block from one track to another. On each track: the eye hides its picture, M mutes it, S solos it, the padlock locks it; a meter shows its level during playback. Right-click a track name to rename it, move it up or down, add a track above or below, duplicate it with its blocks, select all its blocks, change its height (tall, normal, collapsed) and its colour, which its blocks take on, empty or delete it, and remove all empty tracks at once. Double-click its name to rename it, drag its header to move it; the bin in the header deletes it, and the message after a deletion has a Cancel button.
• Edit by text: the Text button, under the preview, lists the sentences spoken in the V1 shots, in order. Click a sentence to go there; the cross removes it from the edit on every track, and the rest moves up. Remove the "uh"s and Remove silences (longer than 0.4, 0.7 or 1.5 s) clean the whole edit in one click. A video that is not transcribed yet can be transcribed from this panel.
• Voice: pick a track (click its name): Reduce hiss removes background noise from its voices, Studio voice also adds a clear voice EQ and a compressor. What you hear in the preview is what gets exported.
• Speed: a video shot plays in slow motion (×0.5, ×0.75) or fast (×1.5, ×2), and its voice keeps its pitch.
• Motion: on a title, an image or a video overlay, Animate adds a first keyframe. Move the playhead, then move or resize the item in the preview: a new keyframe is added, and the item glides from one to the next.
• Cuts on the beat: on a music item, this button moves the cuts between V1 shots onto its beats. Dragged blocks also snap to the beats of background music.
• Save and reopen: the disk button (Ctrl+S) saves the edit to a .autocut-studio file on your computer, the folder button (Ctrl+O) reopens it, here or on another computer. Videos and sounds are not in the file: missing ones are asked for. With the installed app, double-clicking the file opens the Studio.
• Keyboard editing: L plays (twice: faster), J plays backwards, K stops. I and O mark the start and end of a range, X removes it on every track and closes the gap. Holding Ctrl while pulling the right edge of a shot moves the cut with the next one without changing the length of the edit.
• Colour: on a shot, pick a look (warm, cool, punchy, black and white, vintage) or set brightness, contrast, saturation and temperature. Same colour everywhere applies it to every shot, Match cameras corrects the shots from the other cameras so they look like this one.
• Multicam: put the 2nd camera full frame on V2, synced to V1, then click Automatic multicam (video settings): the picture switches between cameras depending on who is speaking.
• Blur faces: finds the faces in the V1 shots and puts a blur zone on each one that follows it. Delete the zone of a face you want to keep.
• Comfort: right-clicking a block offers to split, duplicate, detach its sound or delete it; in an empty part of a track, to paste, close the gap or add a marker. A dragged block snaps to the edges of other blocks, on every track: a dotted line shows it is lined up exactly, and a bubble shows its position. Alt while dragging: no snapping; Shift when dropping: the next blocks on the track are pushed. Alt+← → nudges the selected block by one frame, Z fits the whole edit, F shows the preview full screen and ? lists every shortcut. Several files dropped on the preview go one after another at the end of V1, in name order. Gaps between two blocks on a track are hatched, and the button in the middle closes them. Hover a video in Media: it scrubs under the mouse. The studio reopens where you left it (playhead, zoom), and the clock button keeps a version of the edit every 5 minutes so you can go back. Under the video settings, one tip at a time.
• Drag and drop: drop videos, sounds or images from your computer straight onto a track, at the spot you want, or drag a media from the Media list onto the timeline. On V1 a video becomes a shot; on a video track above, a cutaway; on an audio track, only its sound is kept. Blocks also move from one track to another: a V1 shot dragged to V2 becomes full frame (and back), a video dragged onto an audio track keeps only its sound, and a block released below the last track (or on "Titles and effects" for a video) goes onto a new track. The timeline scrolls when you get near its edge.
• Several blocks: Shift+click adds a block to the selection to move or delete them together. Split all tracks (Shift+S) splits everything under the playhead. Shift+Delete deletes and closes the gap on the track. M adds a marker that blocks snap to.
• Shortcuts: S splits the item under the playhead, Ctrl+C / Ctrl+V copy and paste at the playhead, Ctrl+D duplicates, Delete deletes, Ctrl+Z undoes, Space plays or pauses, ← → step frame by frame, Esc deselects, then closes the studio.
• Transitions: on a shot, choose how it comes in after the previous one (crossfade, fade to black, slide, wipe, zoom) and how long it takes. Same transition everywhere applies it to every shot. The sound blends smoothly during the transition.
• Titles: text, size, colour, style (outline, banner or shadow), entrance animation (fade, pop, slide). Drag it in the preview to place it.
• Overlays and masks: an image or a video on top of the shot, with a shape (rectangle, rounded, round), a border and an opacity. Drag it in the preview, pull a corner to resize it.
• Blur zones: hide a face, a number plate or a screen, blurred or pixelated. The zone stays put: it does not follow a moving object.
• Framing and zooms: for a shot, choose fill, whole or blurred background (the whole picture over a blurred copy, handy to turn a landscape video into a vertical one). A different zoom at the start and the end gives a slow movement. Drag the picture in the preview to reframe it.
• Music: it ducks by itself when someone speaks and comes back up in the silences. Set the volume, how much it ducks and the fades; a song shorter than the video starts over.
• Detached sound: Detach sound, on a shot or a video overlay, moves its sound to an audio track. It stays linked to the picture and follows it when it moves, but you can offset it (sound starting before the picture), trim it or mix it separately. Reattach puts it back with the picture.
• Volume curve: double-click a sound to add a point, drag it to raise or lower the volume there, double-click it to remove it. Two sounds overlapping on the same track crossfade by themselves.
• Mixing: click a track name to adjust it: volume, pan, three-band equalizer (with presets such as Clear voice), compressor. What you hear in the preview is what gets exported. In the video settings, Export loudness brings the mix to -14 LUFS (YouTube, social media) or -16 LUFS (podcast), with a limiter at -1 dB.
• Waveform sync: a separate mic recording or a second camera lines up with the sound of V1 with Sync to the edit's sound, by comparing waveforms.
• Voice-over: the Voice-over button records your mic during playback (headphones keep the video's sound out of your voice). When you stop, the voice lands on an audio track where it started.
• Burned-in captions: in the video settings, on the right when nothing is selected, if the track was transcribed in the Text view. Pick their style (coloured word, banner, growing word, plain), position, size, capitals and words always in colour.
• Export: tick one or more formats (16:9, 9:16, square, 4:5) to get one video per format in one go. The window shows the estimated length and size, and flags what deserves a look first (black screen at the start, title out of frame, clipping sound). A notification tells you when it is done.
• YouTube thumbnail: in the export window, the frame at the playhead with your title in large type and your logo, saved as a 1280 × 720 PNG.

For Shorts:

• Best moments: the Moments button of a video, in Media, listens to the whole track and suggests the liveliest passages (rising voice, highlights, pace, and questions and laughter if the track is transcribed). Choose the length (15, 30 or 60 seconds) and a template, then Edit this moment: the timeline is replaced by that passage, ready to export.
• Templates (at the top): Vertical podcast, Dynamic, Square or Trailer style the whole edit in one click: format, framing, hook title, captions, progress bar, transitions, music. Your cuts don't change, and you can tweak everything afterwards.
• Follow the speaker: in vertical, the frame sits on the person speaking and switches from one to the other, based on movement in the picture (mouth, head, hands). The button is on each shot, or in the video settings for every shot; the Dynamic template does it by itself.
• Blur zones that follow: place the zone on a face or a plate at the playhead, then Follow the movement: it tracks it frame by frame. Drag it afterwards to shift the whole path at once.
• Brand kit (Brand button): your logo and its corner, the title colour, an accent colour (spoken word in captions, progress bar), a font and an end text. It is added to every template, and Apply to this edit puts it on the current edit. It stays in your browser.

Export: choose the format at the top (16:9, 9:16, square or 4:5), then Export the video. The MP4 is written by your browser, with nothing uploaded, in 720p or 1080p. You need Chrome, Edge or a recent Safari. Your edit is kept in the browser and comes back on your next visit, but it is not saved in .autocut files.

Library, map and duplicates

The Library button above the sound list shows every sound of every recording with its thumbnail. Search looks at the name, the recording, the category and what the AI heard. Click a sound to open and play it, click the star (or press F) to favorite it, then filter favorites or export them alone.

Map

The Map button places every sound on a plane: similar sounds (timbre, length, and the kind of sound when the AI is loaded) sit close together, one color per recording. Hover a dot to see its name, click to play it. Search, favorites and tags apply too: filtered-out sounds fade back. Handy to spot gaps in a pack, or sounds that duplicate each other.

Tag

The AI sorts each sound (footsteps, door, water, impact, whoosh, alarm…). Tags show above the thumbnails: click one to see only those sounds.

Duplicates

Finds near-identical sounds across all recordings: the same take exported twice, even at another volume or in another format. Variations of a sound (two different takes) don't count as duplicates. Keep one per group keeps the favorite, otherwise the first sound, and removes the others; Ctrl + Z undoes it.

Mic and generator

Record with the mic

In the recordings panel. The sound comes in raw, without the browser's noise reduction or automatic volume. Make your sounds one after another with a silence between them: when you stop, the recording becomes a new recording and Otocut splits it like a file. The sound never leaves your computer.

Generate blips

Retro sound effects for games and interfaces: click, beep, coin, jump, laser, explosion, hit, powerup, select. Pick a type, click Random until you find the right one (or Variation to change it slightly), adjust the waveform, length, pitch and vibrato, then add it to the list. Create the recording gathers the list into a recording, one sound per blip, already named.

Audacity, Reaper and sound libraries

Import markers

In the Detection tab: a label track exported by Audacity (.txt) or the regions of Reaper's region manager (.csv, with the ruler in minutes:seconds or seconds). Each region becomes a sound, with its name; plain markers split the recording from one to the next. You can also drop the file with its recording: it goes to the recording with the same name. Ctrl + Z undoes it.

Export markers

With Audacity and Reaper markers (Export tab), the .zip contains a markers folder: for each recording, its Audacity labels (File > Import > Labels) and its Reaper regions (region manager, Import).

WAV regions and metadata

A WAV holding named regions (saved by Reaper, Soundminer or another editor) is split along those regions, without detection. The name, UCS category and description stored in a single-sound WAV are kept, the library search reads them, and the export keeps them in its metadata.

Recipes

A recipe keeps all your settings: detection, name template, UCS naming and export (format, sample rate, normalization, variations, engine targets…). Open Recipes in the tracks panel, give it a name and click Save the current settings.

Apply

Puts the studio back into the recipe's settings.

Process a folder

Choose a folder (or files): each file is added, split, named and exported with the recipe, into a single .zip named after the recipe. The recordings stay in the studio in case you want to touch up a sound and export again.

Auto threshold per file

Tick it when saving the recipe: the silence threshold is set from each file's background noise, useful when takes come from different mics or rooms.

Share

The download button creates a .recipe.json file; Import a recipe reads it back, on your computer or a colleague's.

Pack sheet and sound list

Pack sheet

Below the export button: a page with every sound to export, its file name, length, waveform, category and details (loop, favorite, variations), plus the format and total length. Print or save as PDF turns it into a PDF to attach when you deliver a pack.

Sound to-do list

In the tracks panel: paste the list of requested sounds, one per line. Otocut ticks the ones it finds among your sounds (name, category, what the AI heard) and shows which; click a found sound to open it, or an unticked line to tick it yourself. Copy what's missing copies the sounds still to do. The list stays saved in your browser.

Saving and offline

• Your work (recordings, cuts, names) is saved automatically in this browser and comes back on your next visit.
• Save (.autocut) creates a session file, compatible with the Windows app. Reopen it after adding the same recordings.
• Once visited, the studio works without a connection. The Install app button adds it to your computer like an application.

Keyboard shortcuts