{ "categories": { "convert": "Convert a document", "pdf": "PDF page tools", "code": "Code converters", "image": "Images", "svg": "SVG", "audio": "Audio", "other": "Workflow" }, "tools": [ { "slug": "acrobat-for-teachers", "name": "Acrobat for teachers / test-prep creators", "category": "other", "one_liner": "Chat flagship #2", "replaces": "Acrobat + Express generic document tools", "primitive": "the teacher/test-prep workflow as the unit", "status": "planned", "runner": "cli", "script": null, "tests": "", "from": null, "to": null, "public": false }, { "slug": "ai_resume_builder", "name": "AI resume builder", "category": "other", "one_liner": "[mt4] First draft returned 2026-07-10; audited by irp-o.", "replaces": "Adobe/Express resume", "primitive": "schema-first resume + render", "status": "in-progress", "runner": "cli", "script": null, "tests": "", "from": null, "to": null, "public": false }, { "slug": "audio-album-art", "name": "Extract album art", "category": "audio", "one_liner": "Pull the embedded cover art out of any audio file at full resolution.", "replaces": "Screenshotting cover art out of a music player", "primitive": "Pull the embedded artwork stream directly (mutagen first, ffmpeg fallback) at original resolution", "status": "done", "runner": "audiotags", "script": null, "tests": "7/9", "from": null, "to": null, "public": true, "search_title": "Extract album art at full size in its real format", "search_description": "Pull every picture embedded in an audio file out at full resolution, each saved under the format it is actually stored in. Runs locally, nothing is uploaded.", "search_intro": "An audio file can carry more than one embedded picture, and they are not all JPEG. A tool that returns a single file named for the format it expected will hand back PNG bytes under a name ending in jpg, and will drop the second picture without saying so. This page reads the file first and lists what is in it: every attached picture, its format, its pixel size and the label the file gives it. Each one is then copied out under an extension that matches the bytes, with no re-encoding and no resizing.", "how_it_works": [ "Choose an audio file. It is read in this tab and every embedded picture is listed with its format and pixel size.", "Review the pictures, then extract them. Each is copied out at its stored resolution.", "Download them individually. A file with no embedded picture says so instead of producing an empty result." ], "use_cases": [ "Recover cover art from a music library at the resolution the files actually hold.", "Check whether a release carries a back cover or a booklet page as well as a front cover.", "Get artwork out of a podcast or audiobook file before re-tagging it." ], "limits": [ "Only pictures embedded in the file are reachable. Artwork stored beside the file as a separate image is not part of the container.", "Pictures are copied rather than converted, so the format is whatever the file holds and no upscaling happens.", "The role a picture claims, such as front cover against back cover, is not exposed by the container's generic interface and is not reported." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. The file is read and the pictures are copied out in this tab. The audio engine is fetched once and cached." }, { "q": "Why is the extracted file sometimes a PNG?", "a": "Because that is what the file holds. The extension follows the actual stored format rather than being fixed in advance." }, { "q": "What happens with more than one picture?", "a": "Every attached picture is extracted, numbered in the order the file stores them, each with its own format and size." } ], "related": [ "audio-metadata", "audio-analyze", "image-optimizer" ] }, { "slug": "audio-analyze", "name": "Analyze audio", "category": "audio", "one_liner": "Measure integrated loudness, loudness range and true peak against a platform target, and draw the spectrogram and waveform the measurement came from.", "replaces": "Analyzers that make you open a full DAW to see a waveform", "primitive": "One command for waveform, spectrogram, BPM, key, LUFS and peak", "status": "done", "runner": "audioanalyze", "script": null, "tests": "40/40", "from": null, "to": null, "public": true, "search_title": "Measure audio loudness and see the spectrum in the browser", "search_description": "Read integrated loudness in LUFS, loudness range and true peak from a recording, compared against the level your platform normalises to, beside a spectrogram and a waveform drawn from the same file. Runs locally, nothing is uploaded.", "search_intro": "Every consumer platform turns a track up or down at playback until its measured loudness matches a fixed level, so the number that decides how loud a listener hears a recording is calculated after upload and is almost never shown to whoever made it. That number is integrated loudness, expressed in LUFS, and it comes from an algorithm standardised as ITU-R BS.1770 that weights the signal for human hearing and ignores the quiet passages instead of averaging them in. A peak meter cannot answer it: two files can peak at the identical value and sit several LUFS apart, and the quieter one will be turned up while the louder one is turned down. This page measures the loudness figure, the loudness range that says how far the loud and quiet passages sit apart, and the true peak, which is measured between samples rather than only at them and is therefore the one that predicts whether a lossy encoder will clip. Each figure is stated against the target you pick, in text, with the difference in LU. Beside them the same file is drawn twice: a spectrogram on a logarithmic frequency scale, so the musical range is spread evenly rather than squeezed into the bottom of the picture, and a waveform that plots the loudest sample in each column and separates the channels, so a short transient and a lopsided mix are both visible instead of being averaged away.", "how_it_works": [ "Choose one audio file. It is decoded in this tab and its length, sample rate and channel count are read before anything is measured.", "Pick the level your release is aimed at, and the way the pictures should be drawn: logarithmic or linear frequency, peak or average waveform sampling, channels together or apart, and the image size.", "Analyze. Levels are measured in a pass of their own, ahead of the drawing, so a picture that fails to render still leaves the numbers on the page.", "Read the verdict, the loudness table and the per-channel levels, then download either picture or the whole measurement as JSON." ], "use_cases": [ "Find out whether a master will be turned down at playback, and by how much, before it is uploaded anywhere.", "Check a true peak against the ceiling a lossy encoder needs, which a sample peak meter will read as safe when it is not.", "See whether one channel is quieter than the other, or whether a recording has a noise floor or a constant offset, without opening a full editor.", "Look at what actually occupies the top of the spectrum in a recording, on an axis where the musical range is legible." ], "limits": [ "Tempo in beats per minute and musical key are not measured. Both need a library this browser build leaves out, and a guessed answer is worse than no answer.", "The tolerance band used to call a level at target is this tool's own stated band, not a quoted requirement from any standard or platform.", "Loudness is measured over the whole file. A short measurement window and a moving meter are a different job from a delivery check.", "The spectrogram's requested size is the plot area, and the legend is drawn outside it, so the produced picture is larger than the size requested. The size actually produced is read back from the image and reported with it." ], "faq": [ [ "Is the file uploaded?", "No. Decoding, measurement, drawing and export all happen in this browser tab." ], [ "Why does the loudness number differ from my peak meter?", "They measure different things. A peak meter reports the highest sample; integrated loudness weights the signal for hearing, gates out the quiet passages and reports one figure for the whole programme. Two files with identical peaks can differ by several LUFS." ], [ "What is the difference between true peak and peak level?", "Peak level is the loudest sample present. True peak reconstructs the waveform between samples and reports the highest point it reaches there, which can be higher and is what a lossy encoder has to cope with." ], [ "Which target should I pick?", "Pick the one your release is going to. The page states what each level corresponds to, and reports the difference in LU either way, so the measurement is useful whichever target is selected." ] ], "related": [ "audio-normalize", "audio-volume", "audio-dynamics" ] }, { "slug": "audio-censor", "name": "Beep censor", "category": "audio", "one_liner": "Beep out listed time ranges, for profanity or anything you need gone.", "replaces": "Loading a full DAW to bleep three words in a podcast", "primitive": "Time-range list in, beeped copy out; tone frequency and padding are explicit flags", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Beep out parts of an audio recording", "search_description": "Mark the passages to censor on the waveform and replace each with a generated tone. Nothing is uploaded.", "search_intro": "Censoring is a review job before it is an encoding job: the exact passage matters, and a range typed from memory is usually wrong by a syllable. This workspace marks each passage on the waveform, plays it back, and replaces it with a tone generated at the frequency, level and fade chosen here. The surrounding audio is re-cut around the marks and rejoined, and a mark that runs past the end of the recording is refused rather than quietly dropping the tail.", "how_it_works": [ "Choose an audio file and review the waveform.", "Add a beep region for each passage, then drag or type its exact bounds.", "Use play with regions skipped to hear the recording without the censored passages.", "Set the tone frequency, level and fade length.", "Run to build the censored file locally and download it." ], "use_cases": [ "Removing names or numbers from a recorded call before sharing it.", "Bleeping language in a podcast episode.", "Masking a passage in evidence audio without cutting its running time." ], "limits": [ "The tone is synthesized as a sine wave at the chosen frequency, so the censored passage keeps its original length rather than being cut out.", "Regions are rebuilt through a lossless intermediate and re-encoded once at the end.", "A region that starts at or past the end of the recording is refused, and a region ending exactly at the end leaves no trailing segment." ], "faq": [ { "q": "Does the recording get shorter?", "a": "No. Each censored passage is replaced by a tone of the same length, so timings elsewhere are unchanged." }, { "q": "Can the beep be quieter?", "a": "Yes. The level runs from 0 to 1 and the fade softens each end of the tone." }, { "q": "Can the original still be recovered from the output?", "a": "No. The censored passage is replaced before encoding, so the output file contains the tone and not the original audio." } ], "related": [ "audio-silence", "audio-trim", "audio-split" ] }, { "slug": "audio-channels", "name": "Channel tools", "category": "audio", "one_liner": "Convert mono/stereo, swap, isolate or pan channels via an explicit matrix.", "replaces": "Mono/stereo toggles that silently drop a channel", "primitive": "Explicit pan-matrix ops for downmix, upmix, swap, isolate and pan", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-compress", "name": "Compress audio", "category": "audio", "one_liner": "Shrink audio for podcasts and uploads with visible voice/music/maximum tradeoffs.", "replaces": "Compressors that give you one opaque 'compress' button", "primitive": "Named intent presets (voice / music / maximum) that set sample rate, channels and bitrate explicitly", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Shrink audio files and hear what the smaller file lost", "search_description": "Re-encode for podcasts and uploads with visible size reduction and the original on the same transport for comparison. Nothing is uploaded.", "search_intro": "Choosing a bitrate is a trade, and every tool that offers the trade makes it invisible: a smaller file arrives and the question of what it cost goes unanswered. This page reports the new size as a percentage of the original and keeps the untouched recording on the same transport as the re-encoded one, so the bitrate is chosen by listening to both rather than by reputation.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Pick a target profile for voice, music or maximum reduction, or set the bitrate directly.", "Run, then read the size change and compare the two versions." ], "use_cases": [ "Getting a podcast episode under an upload limit without guessing at the bitrate.", "Reducing a set of recordings for archiving where some loss is acceptable.", "Hearing where a particular recording starts to fall apart as the bitrate drops." ], "limits": [ "Lossy re-encoding is not reversible, and re-encoding an already lossy file loses more.", "Speech tolerates far lower bitrates than music; one profile does not suit both.", "A file that is already small may grow, since the encoder targets a bitrate rather than a size." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Why did my file get bigger?", "a": "The chosen bitrate is higher than the source's. Lower it, or keep the original." }, { "q": "How low can I go?", "a": "Compare at each step. Spoken word usually survives much lower settings than music, and the point where it stops working depends on the recording." } ], "related": [ "audio-convert", "audio-normalize", "audio-voice-enhance" ] }, { "slug": "audio-convert", "name": "Audio converter", "category": "audio", "one_liner": "Convert between MP3, WAV, FLAC, AAC, M4A, OGG, Opus, AIFF and WMA with quality-aware, loss-avoiding defaults.", "replaces": "Online audio converters that upload your file, cap size, and re-encode lossy to lossy blindly", "primitive": "Local ffmpeg with a lossy->lossy WAV intermediate so re-encoding never stacks generational loss", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-declick", "name": "Remove clicks and pops", "category": "audio", "one_liner": "Clean clicks and pops off vinyl transfers and damaged recordings.", "replaces": "Restoration suites priced for mastering studios", "primitive": "Three explicit methods (median, lowpass, FFT) with a strength dial instead of one magic button", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Remove clicks and crackle from a vinyl transfer", "search_description": "Take impulse noise off a digitised record or a damaged recording and compare the result against the original before downloading. Nothing is uploaded.", "search_intro": "Clicks are short and loud, which makes them easy to detect and easy to over-treat: a detector set too eagerly starts removing the attack of every drum and consonant, and the recording goes soft in a way that is hard to notice one pass at a time. This page keeps the untreated source on the same transport as the result, so an over-aggressive setting is audible immediately rather than after the file has been filed away.", "how_it_works": [ "Choose a digitised recording. The waveform is drawn locally.", "Pick the method: impulse detection for clicks and pops, smoothing for continuous crackle, or spectral for a mixture.", "Set the strength, run, and compare the result against the source." ], "use_cases": [ "Cleaning a vinyl or shellac transfer before archiving it.", "Removing digital clicks left by a dropout or a bad edit.", "Reducing surface crackle without flattening the music underneath." ], "limits": [ "A strong setting removes transients that belong to the music along with the clicks, which is the trade the comparison exposes.", "Continuous crackle is reduced rather than eliminated, since it overlaps the recording across the whole spectrum.", "Damage longer than an impulse, such as a skipped groove, is not repaired here." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Which method should I use?", "a": "Impulse detection for discrete clicks and pops, smoothing when the noise is continuous crackle, spectral when both are present." }, { "q": "Why does it sound duller?", "a": "The strength is removing transients that belong to the recording. Lower it and compare again." } ], "related": [ "audio-denoise", "audio-dehum", "audio-eq" ] }, { "slug": "audio-dehum", "name": "Remove electrical hum", "category": "audio", "one_liner": "Notch out 50/60 Hz electrical hum and its harmonics without dulling the rest.", "replaces": "One-click denoise that smears the whole spectrum to kill a 60 Hz tone", "primitive": "Cascaded notch filters at the mains frequency and its harmonics, nothing else touched", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Remove mains hum from a recording", "search_description": "Notch out 50 or 60 Hz electrical hum and its harmonics, and hear the result against the original before downloading. Nothing is uploaded.", "search_intro": "Mains hum sits at one frequency and its multiples, which makes it the one noise problem with an exact answer: notch the fundamental and the harmonics and leave everything else alone. The difficulty is that a notch deep enough to kill the hum can also hollow out the low end of a voice, and a downloaded file is a slow way to discover that. This page keeps the source and the result on one transport so the trade is audible in the same screen where the frequency and the harmonic count are chosen.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Pick the mains frequency for the region the recording was made in, 50 Hz or 60 Hz.", "Choose how many harmonics to notch and how deep the notches go.", "Run, then switch the player between source and result to hear what the notches took with them." ], "use_cases": [ "Cleaning a buzz picked up from unbalanced cabling or a ground loop.", "Rescuing an interview recorded next to a fluorescent light or a dimmer.", "Removing hum from a digitised tape without touching the rest of the spectrum." ], "limits": [ "Only hum at the mains fundamental and its harmonics is addressed. Broadband hiss needs the noise reducer instead.", "Deep notches on the fundamental remove real low-frequency content along with the hum, which is why the comparison is on the page.", "Hum that drifts in frequency is only partly caught by fixed notches." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Which frequency should I choose?", "a": "It follows the electricity supply where the recording was made, not where it is being edited." }, { "q": "Why is there still a low rumble?", "a": "Rumble is not hum. It is broadband low-frequency energy and is not at a single frequency, so notches do not remove it." } ], "related": [ "audio-denoise", "audio-eq", "audio-voice-enhance" ] }, { "slug": "audio-denoise", "name": "Reduce noise", "category": "audio", "one_liner": "Reduce steady background noise using a sampled noise profile.", "replaces": "One-click 'remove noise' that guesses the noise profile for you", "primitive": "Sample a real noise clip from the file, then spectral-gate against it (ffmpeg afftdn fallback when the DSP lib is absent)", "status": "done", "runner": "audiosound", "script": null, "tests": "9/11", "from": null, "to": null, "public": true, "search_title": "Reduce background noise with the noise floor measured", "search_description": "Sample a passage that is only noise, see the level it measures, and hear the result against the original before downloading. Nothing is uploaded.", "search_intro": "Noise reduction is a measurement problem wearing the costume of a button. A reducer that assumes a fixed noise floor either scours a clean recording or barely touches a noisy one, and because the only feedback is a downloaded file, nobody finds out which happened. This tool measures the level of a passage that contains nothing but the noise, hands that measurement to the filter as its noise floor, and reports the number on screen. The amount of reduction stays a separate control, because how loud the noise is and how much of it to remove are different questions.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally and nothing is uploaded.", "Point the sample window at a passage that is only room tone, hiss or hum, using the start and length fields.", "Set how many decibels of reduction to apply, then run.", "Read the measured noise floor on the result, and switch the player between source and result to hear the trade." ], "use_cases": [ "Cleaning air-conditioner or computer-fan hiss off an interview recorded in a room.", "Reducing tape or preamp hiss on a digitised cassette or minidisc transfer.", "Checking how much reduction a recording tolerates before the voice starts to sound processed." ], "limits": [ "The sample window must contain only noise. Pointed at speech, it measures the speech and treats it as the floor.", "Spectral reduction removes noise that overlaps the voice by attenuating shared frequencies, so a heavy setting audibly thins the recording. The comparison exists so that is a decision rather than a surprise.", "This is a filter, not a speech model; it does not separate a voice from a competing voice or from music.", "A source with no quiet passage anywhere cannot be measured honestly, and the tool says when the window could not be read instead of assuming a value." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. The measurement, the filtering, the waveforms and the playback all happen in this tab. The audio engine is fetched once and cached." }, { "q": "What does the measured noise floor mean?", "a": "It is the average level of the sample window in decibels, and it is handed to the filter as the level below which audio is treated as noise." }, { "q": "How much reduction should I use?", "a": "Start at the default and raise it until the result begins to sound hollow, comparing against the source each time. There is no correct number for every recording." } ], "related": [ "audio-dehum", "audio-declick", "audio-voice-enhance" ] }, { "slug": "audio-duck", "name": "Voice over music", "category": "audio", "one_liner": "Narration over music with auto-ducking, the podcast mix without the DAW.", "replaces": "Manually riding the music fader under every line of narration", "primitive": "Sidechain auto-ducking: music at full volume until voice is present, then it gets out of the way", "status": "done", "runner": "audiocompose", "script": null, "tests": "6/6", "from": null, "to": null, "public": true, "search_title": "Duck music under narration by a number of decibels", "search_description": "Put a voice over a music bed so the music drops while the voice speaks and returns in the gaps, by a depth you set in decibels. Local, no upload.", "search_intro": "Ducking means the music moves: full level in the gaps, down under the voice, back up afterwards. A control that simply makes the music quiet for the whole piece is not ducking, and it is a common substitute because it is easier to implement and hard to tell apart from a description. This page keeps the two ideas separate. The resting level of the music is one control, and how far the music drops under the voice is another, given in decibels. The depth is reached by measuring how loud the voice actually is and solving the compressor's threshold against that reading, because a fixed threshold produces a different depth on every recording.", "how_it_works": [ "Choose the voice first and the music second. Both appear on one timeline at one shared scale.", "Set how far the music should drop while the voice speaks, and its resting level for the gaps.", "Mix, then read the panel showing the measured voice peak and the threshold that was solved from it." ], "use_cases": [ "Mix a podcast intro where the bed drops under the host and comes back between sentences.", "Put a music bed under a recorded voiceover for a video without editing by hand.", "Produce an audio ad where the bed has to stay audible but never compete with the read." ], "limits": [ "Exactly two files are used, voice then music. A third is refused rather than silently ignored.", "The depth is solved from the peak level of the voice as a whole. A recording whose level varies widely will duck less in its quiet passages than in its loud ones, which is usually what is wanted but is not a fixed number.", "When the voice cannot be measured the page falls back to a documented default and says on the result panel that it did." ], "faq": [ { "q": "Are the files uploaded?", "a": "No. The measurement pass, the mix and the read-back all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Why is the music no longer quiet throughout?", "a": "Because that was not ducking. The music now sits at its own level and drops only while the voice is present, by the depth set on the page." }, { "q": "What is the threshold figure on the result panel?", "a": "It is the level the compressor treats as the point where the voice starts pushing the music down, solved from the measured peak of the voice so that the requested depth is reached." } ], "related": [ "audio-overlay", "audio-voice-enhance", "audio-normalize" ] }, { "slug": "audio-dynamics", "name": "Dynamic range compressor", "category": "audio", "one_liner": "Even out loud and quiet parts with presets or full manual control.", "replaces": "Compressor plugins that assume you own a DAW to host them", "primitive": "FFmpeg acompressor with named presets (vocal, podcast, limiter) and full manual parameter override", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Even out loud and quiet parts of a recording", "search_description": "Compress the dynamic range with presets or full manual control, and compare against the original on one transport. Nothing is uploaded.", "search_intro": "Compression is judged by what it costs, not by what it does. Pulled too far it removes the breathing room that made a performance sound alive, and since it also makes everything louder, the processed version wins a blind comparison for reasons that have nothing to do with quality. This page keeps both versions on one transport and draws both waveforms on a shared scale, so the flattening is visible as a shape even when it is flattering to the ear.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Pick a preset, or set threshold, ratio, attack and release directly.", "Run, then compare the result against the source and look at how much the waveform shape changed." ], "use_cases": [ "Evening out an interview where one speaker is much quieter than the other.", "Controlling peaks before setting a loudness target.", "Learning what a ratio actually does by hearing and seeing it against the original." ], "limits": [ "Compression cannot separate a quiet voice from loud background noise; it raises both.", "Heavy settings pump audibly on material with strong rhythm.", "The result is louder, which biases a listening comparison in its favour; the waveform shape is the check on that." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Should I compress or normalize?", "a": "Compression changes the distance between loud and quiet parts. Normalizing moves the whole recording to a target level and changes nothing else." }, { "q": "What order should these go in?", "a": "Clean the recording first, then compress, then set the loudness target last." } ], "related": [ "audio-normalize", "audio-volume", "audio-eq" ] }, { "slug": "audio-eq", "name": "Parametric EQ", "category": "audio", "one_liner": "Shape the tone with EQ presets or your own multi-band curve.", "replaces": "EQ locked inside player apps that will not export the result", "primitive": "FFmpeg equalizer chains with tonal presets (bass boost, vocal clarity) and manual multi-band config", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Equalize audio with presets or your own bands", "search_description": "Shape the tone with a preset or a typed multi-band curve and hear it against the untouched file. Nothing is uploaded.", "search_intro": "Equalization is the one audio edit where a small change can be inaudible and a large one irreversible, and where the ear adapts within seconds of hearing only the result. The defence is an immediate comparison: this page keeps the original and the equalized version on one transport, and draws both waveforms on a single shared scale so a change in level is visible as well as audible.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Pick a preset, or type a band list to override it with exact frequencies, widths and gains.", "Run, then switch between source and result." ], "use_cases": [ "Lifting the presence range so a voice cuts through on small speakers.", "Rolling off rumble below the useful content of a recording.", "Applying an exact, repeatable band list rather than dragging an unlabelled curve." ], "limits": [ "Boosting a frequency that carries no content amplifies the noise there instead.", "The band list is parsed strictly; a band missing a frequency, width or gain is refused rather than guessed at.", "Output is re-encoded, so a lossy source stays lossy." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "What does the band syntax mean?", "a": "Each band names a centre frequency, a filter type, a width and a gain in decibels, and bands are separated by a vertical bar." }, { "q": "Why did the level change?", "a": "Boosting bands raises overall level. The two waveforms share one scale so the change is visible, and the loudness tool sets a target afterwards." } ], "related": [ "audio-dynamics", "audio-normalize", "audio-voice-enhance" ] }, { "slug": "audio-extract-from-video", "name": "Extract audio from video", "category": "audio", "one_liner": "Pull the audio track out of MP4, MKV, MOV, WebM and more into any audio format.", "replaces": "'Extract audio' web tools that upload the whole video just to strip the track", "primitive": "Decode only the audio stream locally; the video never leaves the machine", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-karaoke", "name": "Basic vocal remover", "category": "audio", "one_liner": "Remove center-panned vocals via phase cancellation (basic, honest about its limits).", "replaces": "'Remove vocals' tools that promise AI stems and deliver mush", "primitive": "Honest center-channel phase cancellation, labeled basic rather than sold as stem separation", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Remove center-panned vocals by phase cancellation", "search_description": "Cancel what is common to both channels and hear exactly what that costs, against the original. Honest about what it cannot do. Nothing is uploaded.", "search_intro": "Phase cancellation removes whatever sits equally in both channels. On most mixes that is the lead vocal, and also the kick drum, the bass and anything else placed in the middle. It is not vocal separation and no amount of tuning turns it into vocal separation. The useful thing a page can do is make the cost audible in one click, which is what the comparison here is for.", "how_it_works": [ "Choose a stereo file. A mono file has nothing to cancel against and is not a candidate.", "Pick the method and run.", "Compare the result against the source to hear what else disappeared with the vocal." ], "use_cases": [ "Making a rough backing track from a stereo mix.", "Checking quickly whether a particular mix responds to cancellation at all.", "Hearing what is centre-panned in a mix." ], "limits": [ "Bass, kick and any other centred instrument is removed along with the vocal.", "A vocal with stereo reverb or double-tracking leaves audible residue.", "A mono source cannot be processed this way at all.", "This is phase cancellation, not source separation, and it does not use a model." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Why does the bass disappear too?", "a": "Bass is usually centred, and cancellation removes what is common to both channels regardless of what it is." }, { "q": "Can this isolate the vocal instead?", "a": "No. This method removes what is centred; it cannot produce it on its own." } ], "related": [ "audio-channels", "audio-eq", "audio-dynamics" ] }, { "slug": "audio-loop", "name": "Loop audio", "category": "audio", "one_liner": "Loop a file N times or out to a target duration, seamless if you want.", "replaces": "Copy-pasting a track four times on a timeline to fill two minutes", "primitive": "Loop N times or to a target duration, with optional crossfade at the seams", "status": "done", "runner": "audiocompose", "script": null, "tests": "8/8", "from": null, "to": null, "public": true, "search_title": "Loop audio to a length that is actually the length you asked for", "search_description": "Repeat a clip a set number of times or out to a target duration, cut to that duration exactly, with an optional crossfade at each join. Local, no upload.", "search_intro": "A target duration is a promise about the finished file, and a looper that repeats until it has passed the target and then stops has not kept it. Asking for sixty seconds from a seven second clip gives sixty three, and adding a crossfade makes the error change sign, because each overlap removes time the repeat count never accounted for. This page counts the repeats with the overlap included and then cuts the result to the target, so the number on the page and the length of the file are the same number. Whole repetitions are still available as an explicit choice rather than as the only behaviour.", "how_it_works": [ "Choose one file. It is decoded in this tab and drawn against the timeline the current settings imply.", "Loop by a repeat count, or by a target duration and let the page work out the repeats.", "Render, then compare the predicted length against the length read back from the produced file." ], "use_cases": [ "Extend a short music bed to the exact runtime of a video.", "Turn a loop into an hour of background audio for a stream or a workspace.", "Repeat a test tone or a metronome bar out to a fixed practice length." ], "limits": [ "Cutting to an exact target re-encodes, because a stream copy can only cut on a frame boundary. The page says which of the two happened and why.", "A stream copy of a lossy source carries an encoder padding frame into every join, so the seams are not sample accurate. The page reports the padding per join rather than claiming a seamless result.", "Whether a loop is audible at the join depends on how the source starts and ends. No looper can make an arbitrary clip meet itself cleanly." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Decoding, looping and the read-back all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Why did a sixty second target used to give sixty three seconds?", "a": "The repeat count was rounded up and the result was never cut. It is now cut to the target, and the whole-repetition behaviour is a checkbox." }, { "q": "Does a crossfade change how many repeats are needed?", "a": "Yes. Every join gives up the crossfade length, so more repeats are needed to reach the same target. That is now part of the count." } ], "related": [ "audio-merge", "audio-trim", "audio-ringtone" ] }, { "slug": "audio-merge", "name": "Merge audio", "category": "audio", "one_liner": "Join multiple audio files into one, with an optional crossfade.", "replaces": "Joiners that glue mismatched files and produce clicks at the seams", "primitive": "Normalize every input to a common format first, then concat (optionally crossfade)", "status": "done", "runner": "audiocompose", "script": null, "tests": "5/5", "from": null, "to": null, "public": true, "search_title": "Join audio files in an order you can see and change", "search_description": "Put several recordings end to end, with the order on a timeline and an optional crossfade at each join. Runs locally, nothing is uploaded.", "search_intro": "Joining files is an ordering job, and most joiners treat the order as whatever the file picker handed over. This page puts every file on one timeline at one shared amplitude scale, so a track that is quieter than the rest is visible before anything is rendered, and the order is moved with buttons rather than only by dragging. A crossfade overlaps each join, which makes the result shorter than the sum of the parts by one overlap per join; the page states the length the settings predict and then, after the render, the length the finished file actually has.", "how_it_works": [ "Choose two or more files. Each one is decoded in this tab and drawn as a track on the shared timeline.", "Move tracks up and down until the order is the one you want, and set a crossfade if the joins should overlap.", "Join them, then compare the predicted length against the length read back from the produced file." ], "use_cases": [ "Assemble interview segments or takes into one continuous recording.", "Build a continuous set from separate tracks with a crossfade at each change.", "Concatenate chapters into a single long file for a player that has no queue." ], "limits": [ "A crossfade at or past the whole length of the shortest file is refused by name, because the overlap would consume that file entirely and the result would be shorter than the inputs.", "Every input is converted to a common sample rate and channel count before joining, taking the highest of each, so a mono file in a stereo set is upmixed.", "Trimming inside a track belongs to the trim tool; this page joins whole files." ], "faq": [ { "q": "Are the files uploaded?", "a": "No. Decoding, drawing and joining all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Why is the result shorter than the sum of the files?", "a": "A crossfade overlaps the end of one file with the start of the next, so each join gives up that much time. With no crossfade the result is the exact sum." }, { "q": "Can the order be changed after adding files?", "a": "Yes. Each track has up and down controls, and the timeline redraws in the new order." } ], "related": [ "audio-overlay", "audio-loop", "audio-trim" ] }, { "slug": "audio-metadata", "name": "Edit metadata", "category": "audio", "one_liner": "Set title, artist, album, year, genre and track tags.", "replaces": "Tag editors locked to a single container format", "primitive": "One tag interface over ID3, Vorbis comments and MP4 atoms", "status": "done", "runner": "audiotags", "script": null, "tests": "7/9", "from": null, "to": null, "public": true, "search_title": "Edit audio tags after seeing the ones already there", "search_description": "Read the title, artist, album and every other tag a file already carries, change them, remove them, and read the written file back to check. Local, no upload.", "search_intro": "Most tag editors present empty boxes and ask what the tags should be, which means retyping values the file already holds and guessing at whatever is not shown. This one opens by reading the file: each field arrives filled in with what the container actually says, every embedded picture is listed at its real format and pixel size, and the facts a container reports about itself are shown without being offered for editing. The written file is then reopened and read a second time, so what appears on the page after a write comes from the produced file rather than from the request that made it.", "how_it_works": [ "Choose one audio file or a batch. Each is read in this tab and its current tags appear in the fields.", "Edit any field, tick remove beside a tag to drop it, or clear every tag at once. Artwork is a stream rather than a tag and is kept either way.", "Write the file, then read the result panel, which is filled from a second read of the file that was just produced." ], "use_cases": [ "Correct the artist or album on files ripped or downloaded with the wrong tags.", "Strip personal comments, encoder notes and stray tags before sharing a recording.", "Check what a file actually carries before sending it to a distributor or a player that reads tags strictly." ], "limits": [ "Tags are written through the container's generic key and value interface, so lyrics, chapters, ratings and replay gain are out of reach.", "MP3 output is written as ID3 version 2.3, which more players and file browsers read than version 2.4. ID3 version 1 is not written alongside it.", "The picture role byte that distinguishes a front cover from a back cover is not reachable, so a written cover is simply a cover." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Reading the tags, writing the new file and reading it back all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Does editing a tag re-encode the audio?", "a": "No. Every stream is copied, so the audio and any embedded artwork come through untouched and only the tags are rewritten." }, { "q": "Why does an empty field not delete the tag?", "a": "An empty field cannot say whether it was cleared or never filled. Removal is a separate instruction, which is what the remove box beside each field does." } ], "related": [ "audio-album-art", "audio-convert", "audio-analyze" ] }, { "slug": "audio-normalize", "name": "Normalize loudness", "category": "audio", "one_liner": "Normalize loudness to a target LUFS (or peak/RMS) via a measured two-pass pass.", "replaces": "'Normalize' buttons that only peak-normalize", "primitive": "Broadcast-standard two-pass EBU R128 loudnorm (LUFS), with peak and RMS as explicit alternatives", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Normalize loudness to a LUFS target with the measurement shown", "search_description": "Measure the integrated loudness, true peak and range of a recording, then move it to a target. The measured numbers are reported, not just the target. Nothing is uploaded.", "search_intro": "Every service that plays audio back turns it down to its own loudness target, so a file mastered louder than that target is not louder on arrival, only more compressed on the way. Getting this right means knowing two numbers: where the recording sits and where it should sit. Most browser normalizers show neither and return a file. This one runs a real loudness analysis, reports the integrated loudness, true peak and loudness range it measured, and then applies the correction to the target chosen here.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Pick the method: integrated loudness for anything destined for a streaming service, or peak and average level for simpler cases.", "Set the target and run.", "Read the measured loudness of the input beside the target, and compare the result against the source." ], "use_cases": [ "Bringing a podcast episode to a consistent level across a series.", "Finding out where a recording actually sits before deciding whether it needs anything at all.", "Matching a set of files to one target so a playlist does not jump between tracks." ], "limits": [ "Integrated loudness is a measurement over a whole recording; a short clip or a clip that is mostly silence produces a figure that is technically correct and practically misleading.", "Where the correction would push the true peak past its ceiling, the loudness filter falls back to a dynamic mode, so the reported measurement describes the input rather than guaranteeing the output.", "Loudness normalization does not repair clipping that is already in the recording." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. The analysis and the correction both run in this tab." }, { "q": "What target should I use?", "a": "Streaming services publish their own; the default here is the common one for speech and music delivery. The right answer depends on where the file is going." }, { "q": "Why is the measured number different from the target after processing?", "a": "The reported measurement is of the input. It is shown so the size of the correction is visible rather than hidden." } ], "related": [ "audio-volume", "audio-dynamics", "audio-compress" ] }, { "slug": "audio-overlay", "name": "Overlay two tracks", "category": "audio", "one_liner": "Mix one track over another with volume, delay and fade control.", "replaces": "Opening an editor just to lay one sound over another", "primitive": "Two files in, one mix out; per-track volume, overlay delay and fade-in as flags", "status": "done", "runner": "audiocompose", "script": null, "tests": "7/7", "from": null, "to": null, "public": true, "search_title": "Lay one track over another without the base dropping", "search_description": "Mix an overlay onto a base track with independent levels, a start offset and a fade, and keep the base at the level you set for the whole file. Local, no upload.", "search_intro": "Mixing two tracks is a level job, and the usual mixing filter quietly rescales its inputs by how many of them are currently playing. The result is a base that sits six decibels down for exactly as long as the overlay runs and jumps back the instant it ends, which is heard as a step in the middle of the mix rather than as a setting anyone chose. This page turns that rescaling off, so each level is the level asked for, and then says what the two peaks sum to, because summing without a rescale is what makes clipping possible. A limiter is offered as a decision rather than applied as a silent correction.", "how_it_works": [ "Choose the base first and the overlay second. Both appear on one timeline, the overlay drawn at the offset you set.", "Set each level, the start offset and any fade-in, and read the worst-case summed peak the page reports.", "Mix, then check the receipt read back from the produced file." ], "use_cases": [ "Put a music bed under narration that is already recorded.", "Drop a sound effect at a known second inside a longer recording.", "Layer an ambience track under dialogue without either being rescaled." ], "limits": [ "Exactly two files are used. A third is refused rather than silently ignored, which is what happened before.", "The two levels are summed as set, so a loud pair can pass full scale. The page reports the worst case and offers a limiter; it does not quietly turn anything down.", "For narration that should push the music down only while it speaks, the ducking tool is the right one; this page holds both levels constant." ], "faq": [ { "q": "Are the files uploaded?", "a": "No. Decoding, mixing and the read-back all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Why did my base used to get quieter and then louder again?", "a": "The mixing filter normalizes by the number of active inputs by default, so the base was halved while the overlay played. That normalization is now off." }, { "q": "What does the summed peak figure mean?", "a": "It is the worst case, the level reached if both peaks landed on the same sample. Real material usually peaks lower." } ], "related": [ "audio-duck", "audio-merge", "audio-volume" ] }, { "slug": "audio-pitch", "name": "Change pitch", "category": "audio", "one_liner": "Shift pitch by semitones. Tempo shifts with pitch.", "replaces": "Pitch shifters that drag the tempo along with the pitch", "primitive": "Semitone-based shift with an optional tempo-preserving mode and an asetrate fallback", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-rearrange", "name": "Rearrange clips", "category": "audio", "one_liner": "Cut listed segments out of a file and rejoin them in any order.", "replaces": "Razor-blade editing for a simple reorder of known segments", "primitive": "A range list like 0-15,30-45,15-30 extracts and concatenates in that order, crossfade optional", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Reorder sections of an audio file", "search_description": "Pick clips on the waveform, put them in the order you want, and export one joined file. Nothing is uploaded.", "search_intro": "Reordering is the one audio job where the list order is the output, and a comma separated string of ranges hides that completely. This workspace makes each clip a row that can be moved earlier or later, with the waveform showing where every clip comes from. The clips are extracted through a lossless intermediate and joined in list order, optionally with a crossfade, so the exported file follows the arrangement on screen.", "how_it_works": [ "Choose an audio file and review the waveform.", "Add one clip per section and set its bounds by dragging, by keyboard, or by typing.", "Move clips earlier or later until the row order matches the arrangement wanted.", "Set a crossfade if the joins should be smoothed.", "Run to build the joined file locally and download it." ], "limits": [ "Clip order is the output order; the list is never sorted by time, which is what separates this from every other region tool here.", "A crossfade longer than the shortest clip is reduced to half that clip's length.", "The same passage can appear more than once, and each appearance is extracted separately." ], "use_cases": [ "Moving an answer before its question in an interview edit.", "Assembling a highlight reel from one long recording.", "Building a rough arrangement from takes inside a single file." ], "faq": [ { "q": "Can a clip be used twice?", "a": "Yes. Add it as two rows; each is extracted and placed independently." }, { "q": "Does the order follow the waveform?", "a": "No. The row order is the output order, so a later passage can be placed first." }, { "q": "What does the crossfade do at the joins?", "a": "It overlaps each pair of neighbouring clips for the chosen number of seconds instead of butting them together." } ], "related": [ "audio-split", "audio-trim", "audio-merge" ] }, { "slug": "audio-reverse", "name": "Reverse audio", "category": "audio", "one_liner": "Play it backwards. Format preserved unless you say otherwise.", "replaces": "Web toys that upload your file to play it backwards", "primitive": "FFmpeg areverse, format preserved by default, re-encode only if you ask", "status": "done", "runner": "audiotool", "script": null, "tests": "", "from": null, "to": null, "public": false }, { "slug": "audio-ringtone", "name": "Make a ringtone", "category": "audio", "one_liner": "Cut a 30s iPhone (M4R) or Android (MP3) ringtone with fades.", "replaces": "Ringtone makers that upsell the export or the length", "primitive": "Length-clamped trim+fade encoded straight to M4R (iPhone) or MP3 (Android)", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Make a ringtone from any audio file", "search_description": "Choose the hook on a waveform and export it as an iPhone .m4r or an Android .mp3, inside the length the phone accepts. Nothing is uploaded.", "search_intro": "The part of a song that works as a ringtone is a few seconds long and almost never where a typed start time lands. This workspace draws the waveform, lets the selection be dragged or nudged by keyboard, and plays it back before encoding. It also enforces the one constraint that silently breaks ringtones: iOS refuses to sync an .m4r longer than 40 seconds, so a longer selection is refused here with the reason instead of producing a file the phone ignores.", "how_it_works": [ "Choose an audio file and review the waveform.", "Drag the selection over the passage wanted, or nudge either edge with the arrow keys.", "Play the selection to check where it starts and ends.", "Pick iPhone or Android, add fades if the cut is abrupt, and run.", "Download the .m4r or .mp3 and move it to the phone." ], "use_cases": [ "Cutting a ringtone out of a track already owned.", "Making a short alert tone from a recording.", "Producing the same hook in both phone formats." ], "limits": [ "An iPhone ringtone must be 40 seconds or shorter to sync; a longer selection is refused rather than encoded.", "The iPhone output is AAC in an ipod container with the .m4r extension; Android output is MP3.", "Installing the file on the phone is a manual step this tool does not perform." ], "faq": [ { "q": "Why is 40 seconds the limit?", "a": "It is Apple's constraint. An .m4r longer than 40 seconds will not sync, so it is refused here rather than produced." }, { "q": "How do I install it on an iPhone?", "a": "Transfer the .m4r to the device through Finder or iTunes; it then appears under ringtones in the sound settings." }, { "q": "Does Android have the same limit?", "a": "No. The Android MP3 output has no length cap here." } ], "related": [ "audio-trim", "audio-split", "audio-convert" ] }, { "slug": "audio-silence", "name": "Remove or insert silence", "category": "audio", "one_liner": "Remove or insert silence with explicit threshold and duration controls.", "replaces": "Silence removers stuck on one fixed threshold", "primitive": "Explicit threshold + min-duration for removal, and position + duration for insertion", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Remove silence from a recording, after reviewing it", "search_description": "See the detected silences on the waveform, adjust or delete any of them, then rebuild the file from what is left. Nothing is uploaded.", "search_intro": "A silence threshold applied without review is the fastest way to lose a quiet word. This workspace detects the silent passages, draws them on the waveform, and lets every one of them be adjusted or removed from the list before anything is cut. Detection runs two ways and says which one produced the current regions: an instant browser estimate, and a measured pass by the audio engine that performs the cut. The export rebuilds the file from the material between the reviewed regions, so what is removed is exactly what was shown.", "how_it_works": [ "Choose an audio file; an estimate of the silent passages appears on the waveform immediately.", "Adjust the threshold and minimum length, then re-estimate, or measure with the audio engine for its own answer.", "Edit or delete individual regions, and use play with regions skipped to hear the result first.", "Run to rebuild the file from the kept material and download it." ], "use_cases": [ "Tightening the pauses in a podcast or interview.", "Trimming dead air from the front and back of a voice memo.", "Cleaning up gaps in a lecture recording before publishing." ], "limits": [ "The regions shown on load are a browser-side estimate; the measured pass comes from the audio engine's own silencedetect and is labelled as such.", "Silence running to the end of the file has no detected end point, so it is closed at the source length.", "Regions covering the entire file are refused, since nothing would remain." ], "faq": [ { "q": "What is the difference between estimated and measured?", "a": "The estimate is computed from the decoded waveform in this tab and is instant. The measured pass runs the audio engine's own detection, which is the same engine that performs the cut." }, { "q": "Can I keep a pause the detector flagged?", "a": "Yes. Remove that region from the list and it stays in the output." }, { "q": "Can silence be added instead?", "a": "The local command line tool has an insert mode for placing a gap at a chosen position." } ], "related": [ "audio-trim", "audio-split", "audio-censor" ] }, { "slug": "audio-speed", "name": "Change speed", "category": "audio", "one_liner": "Change playback speed. Pitch shifts with speed.", "replaces": "Speed changers that chipmunk the pitch up with the tempo", "primitive": "atempo with an optional pitch-preserving mode", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-split", "name": "Split audio", "category": "audio", "one_liner": "Split one file into many by duration, segment count, or detected silence.", "replaces": "Splitters that only cut by a single fixed duration", "primitive": "Three explicit split modes: equal duration, N segments, or silence-detected gaps", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Split audio at cut points you can see", "search_description": "Place cut points on the waveform and export the segments as separate files, numbered in order. Nothing is uploaded.", "search_intro": "Splitting by a chunk length is a guess about where the music or the speech actually breaks. This workspace puts the cut points on the waveform, so each boundary and the segment it opens can be reviewed and played before export. Every segment is written as its own file, numbered in source order, and the run is refused outright when the length of the source cannot be read, which is the case that previously produced silently empty output.", "how_it_works": [ "Choose an audio file and review the drawn waveform.", "Drag the cut point, add more with the Add cut point control, or type exact times.", "Play from any cut point to hear where the boundary falls.", "Pick an output format and run; each segment is encoded locally.", "Download the segments individually or as one ZIP." ], "use_cases": [ "Breaking a long recording into per-track or per-topic files.", "Splitting a lecture at the points where the subject changes.", "Cutting a DJ set into individual pieces at the transitions." ], "limits": [ "Segment files are named with the source name plus a three digit index, so ordering survives a file listing.", "Cut points closer than 0.05 seconds apart are refused, because they would produce a segment shorter than one encoder frame.", "A source whose length the browser cannot read is refused instead of split, since every boundary depends on that length." ], "faq": [ { "q": "How are the output files named?", "a": "The source name, an underscore, and a zero padded index: recording_001.mp3, recording_002.mp3, and so on." }, { "q": "Can I still split by chunk length?", "a": "The local command line tool keeps the chunk length, equal segments and silence gap modes. The browser workspace uses reviewed cut points so the picture matches the result." }, { "q": "Is there a limit on the number of segments?", "a": "No fixed limit, but every segment is encoded in this tab, so a very large number takes proportionally longer." } ], "related": [ "audio-trim", "audio-silence", "audio-rearrange" ] }, { "slug": "audio-tone", "name": "Tone and noise generator", "category": "audio", "one_liner": "Generate test tones and noise at a chosen frequency, duration and level.", "replaces": "Ad-laden tone-generator websites", "primitive": "Local generation of sine/square/saw tones and white/pink/brown noise at exact parameters", "status": "done", "runner": "audiotool", "script": null, "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "audio-trim", "name": "Trim audio", "category": "audio", "one_liner": "Cut audio to a start/end range with optional fade in and out.", "replaces": "Trimmers that re-encode the whole clip to cut two seconds", "primitive": "Timestamp-accurate atrim with optional fades, keeping the codec where possible", "status": "done", "runner": "audioregions", "script": null, "tests": "7/7", "from": null, "to": null, "public": false, "search_title": "Trim audio on a waveform, in the browser", "search_description": "Cut a clip to a start and end you can see and hear, with fades and a remove-the-middle mode. Nothing is uploaded.", "search_intro": "A timestamp typed into a box is a guess. This workspace draws the waveform, puts the start and end on it as draggable, keyboard-operable handles, and plays the selection back before anything is encoded. The cut it performs is the range it drew: the same region array produces the picture and the engine arguments. A range that cannot produce a real file, such as an end before its start or a start past the end of the recording, is refused with the reason rather than handed to the encoder.", "how_it_works": [ "Choose an audio file; its length is read from the browser's own media decoder and the waveform is drawn locally.", "Drag either edge of the selection, or focus an edge and use the arrow keys, Home and End.", "Play the selection to hear exactly what will be kept, then set fades if wanted.", "Switch on remove-the-middle to keep the outside of the selection instead of the inside.", "Run to encode the result locally and download it." ], "use_cases": [ "Cutting an intro and outro off a recording.", "Pulling a single quote out of an interview.", "Removing a section from the middle of a take without opening a full editor." ], "limits": [ "Remove-the-middle rebuilds the file from the surviving sides through a lossless intermediate; a selection covering the whole file is refused.", "The waveform is drawn only for clips short enough to decode in a browser; longer clips keep numeric and playhead-driven region entry.", "The output is re-encoded in the chosen format, so a lossy source stays lossy." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Decoding, the waveform, playback and the encode all happen in this tab. The audio engine is fetched once and cached." }, { "q": "Why was my range refused?", "a": "A range whose end is at or before its start, or whose start is past the end of the recording, cannot produce audio. The message names which one applies." }, { "q": "Can I keep the outside instead of the inside?", "a": "Yes. Remove-the-middle keeps everything outside the selection and joins the two sides." } ], "related": [ "audio-split", "audio-ringtone", "audio-silence" ] }, { "slug": "audio-voice-enhance", "name": "Enhance voice", "category": "audio", "one_liner": "Clean up spoken-word audio with a podcast/broadcast/phone filter chain.", "replaces": "'Enhance voice' AI toggles you cannot see or adjust", "primitive": "A named EQ+compression chain (podcast/broadcast/phone) you can read and tweak", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Clean up spoken-word audio and hear the difference", "search_description": "Apply a podcast, broadcast or phone voice chain and compare it against the untouched recording on one transport. Nothing is uploaded.", "search_intro": "A voice chain is several decisions at once: what to cut below the voice, what to lift in the presence range, and how hard to even out the level. Presets make those decisions quickly and hide them completely, so the only way to judge one is to hear it against the recording it replaced. This page puts both on the same transport and reports the size of what came out, so choosing a chain is a comparison rather than a guess.", "how_it_works": [ "Choose a spoken-word recording. The waveform is drawn locally.", "Pick a chain: podcast, broadcast, or phone.", "Run, then switch between source and result to hear what the chain did." ], "use_cases": [ "Preparing an interview for a podcast without opening a full editor.", "Making a lecture recording intelligible on laptop speakers.", "Deciding which chain suits a particular voice by hearing all of them against the original." ], "limits": [ "A chain is a fixed set of decisions and does not adapt to the recording. A voice it does not suit sounds worse, which the comparison shows.", "Enhancement cannot recover detail a lossy encoder or a clipped preamp already discarded.", "Background noise is shaped by these filters but not removed; the noise reducer is the tool for that." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Should I reduce noise before or after this?", "a": "Reduce noise first, then send the result straight into this tool with the button that makes a result the new source." }, { "q": "Why does the phone chain sound thin?", "a": "It is meant to. It emulates a narrowband telephone path and is there for that effect." } ], "related": [ "audio-denoise", "audio-dynamics", "audio-normalize" ] }, { "slug": "audio-volume", "name": "Change volume", "category": "audio", "one_liner": "Raise or lower volume by an exact dB or a linear factor.", "replaces": "Volume sliders with no numbers behind them", "primitive": "Exact dB gain or linear level, applied in a single filter pass", "status": "done", "runner": "audiosound", "script": null, "tests": "4/6", "from": null, "to": null, "public": true, "search_title": "Change volume by an exact amount", "search_description": "Raise or lower level by a stated number of decibels or a linear factor, with an optional limiter, and see the change in the waveform. Nothing is uploaded.", "search_intro": "A volume change is the one audio edit with an exact answer, and it is still worth seeing. Raising the level of a recording that is already close to the ceiling clips it, and clipping is far easier to see in a waveform than to hear on the first pass. This page draws the source and the result on one shared scale, so the change is a visible change in height rather than two pictures that look the same because each was drawn to fit its own box.", "how_it_works": [ "Choose an audio file. The waveform is drawn locally.", "Set the change as decibels of gain, or as a linear factor.", "Switch on the limiter if the level is going up and headroom is uncertain.", "Run, then compare the two waveforms and the two versions." ], "use_cases": [ "Lifting a recording that was captured too quietly.", "Pulling a file down by an exact amount before mixing it with another.", "Seeing how close to the ceiling a recording already sits." ], "limits": [ "Raising the level raises the noise with it; nothing here separates them.", "Without the limiter, a large increase clips. With it, the peaks are held back and the shape changes.", "A recording already clipped in the source is not repaired by turning it down." ], "faq": [ { "q": "Is the file uploaded?", "a": "No. Everything runs in this tab." }, { "q": "Should I use this or the loudness tool?", "a": "This applies an exact change chosen here. The loudness tool measures the recording and works out the change needed to reach a target." }, { "q": "Why do both waveforms use the same scale?", "a": "So that a level change is visible. Drawing each one to fill its own box would hide exactly the thing being changed." } ], "related": [ "audio-normalize", "audio-dynamics", "audio-compress" ] }, { "slug": "batch-renamer", "name": "Batch renamer", "category": "image", "one_liner": "Rename files as a previewable validated plan: dry-run by default, two-phase cycle-safe apply, undo journal.", "replaces": "Bridge batch rename", "primitive": "Rename as a previewable plan (mandatory dry-run), not a blind find-replace", "status": "done", "runner": "imgtool", "script": null, "tests": "18/18", "from": null, "to": null, "public": false }, { "slug": "catalog-generator", "name": "Catalog generator", "category": "image", "one_liner": "Generate a catalog/price-sheet as data-driven output from a CSV (the data is the source of truth).", "replaces": "InDesign/Express catalog + price-sheet layout", "primitive": "Catalog as data-driven output from a CSV (the data is the source of truth), not a hand-laid document", "status": "done", "runner": "imgtool", "script": null, "tests": "11/11", "from": null, "to": null, "public": false }, { "slug": "cc-commonjs-to-esm", "name": "commonjs-to-esm", "category": "code", "one_liner": "Converts CommonJS modules to ESM: require() → import, module.exports → export, exports.x → export const x, __dirname/__filename → import.meta.url helpers.", "replaces": "Manual require/import rewrites and copy-paste markdown conversion", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "module/commonjs-to-esm.js", "tests": "1/1", "from": "commonjs", "to": "esm", "public": false }, { "slug": "cc-cra-to-vite", "name": "cra-to-vite", "category": "code", "one_liner": "Migrates Create React App (CRA) projects to Vite. Generates vite.config.js, updates env vars (REACT_APP_ → VITE_), and suggests package.json script changes.", "replaces": "Tool migration guides followed by hand, file by file", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "cli", "script": null, "tests": "1/1", "from": "cra", "to": "vite", "public": false }, { "slug": "cc-css-px-to-clamp", "name": "css-px-to-clamp", "category": "code", "one_liner": "Converts px font-size, padding, margin, and gap values to CSS clamp() for fluid responsive sizing. Uses viewport-width-based preferred value with rem min/max.", "replaces": "Tool migration guides followed by hand, file by file", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "config/css-px-to-clamp.js", "tests": "1/1", "from": "css-px", "to": "clamp", "public": false }, { "slug": "cc-css-px-to-rem", "name": "css-px-to-rem", "category": "code", "one_liner": "Converts px units to rem in CSS. Preserves 0px as 0, optionally keeps 1px for borders, handles calc() expressions, var(), @media queries, @supports, and !important.", "replaces": "Tool migration guides followed by hand, file by file", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "config/css-px-to-rem.js", "tests": "1/1", "from": "css-px", "to": "rem", "public": false }, { "slug": "cc-css-to-css-modules", "name": "css-to-css-modules", "category": "code", "one_liner": "Converts global CSS to CSS Modules format. Class selectors kept as-is (local by default). Element/tag and ID selectors converted to class equivalents with warnings. Global resets (html, body) wrapped in :global(). Generates optional JS import map with --with-map flag. Suggests moving @import rules to JS imports.", "replaces": "Manual CSS refactors and single-purpose style converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "css/css-to-css-modules.js", "tests": "1/1", "from": "css", "to": "css-modules", "public": false }, { "slug": "cc-css-to-emotion", "name": "css-to-emotion", "category": "code", "one_liner": "Converts CSS rules to Emotion-compatible syntax. Supports css prop mode (--mode=prop) and styled mode (--mode=styled). Handles @media nesting, keyframes, and CSS variable → theme mapping.", "replaces": "Manual CSS refactors and single-purpose style converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "css/css-to-emotion.js", "tests": "1/1", "from": "css", "to": "emotion", "public": false }, { "slug": "cc-css-to-scss", "name": "css-to-scss", "category": "code", "one_liner": "Converts flat CSS into nested SCSS by grouping related selectors, extracting repeated values to variables, and detecting BEM patterns. Simple nesting heuristics: shared prefix selectors get nested.", "replaces": "Manual CSS refactors and single-purpose style converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "css/css-to-scss.js", "tests": "1/1", "from": "css", "to": "scss", "public": false }, { "slug": "cc-css-to-styled-components", "name": "css-to-styled-components", "category": "code", "one_liner": "Converts CSS rules to styled-components (React) syntax. Selector names become PascalCase component names. Nested selectors use & references or generate separate components.", "replaces": "Manual CSS refactors and single-purpose style converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "css/css-to-styled-components.js", "tests": "1/1", "from": "css", "to": "styled-components", "public": false }, { "slug": "cc-css-to-tailwind", "name": "css-to-tailwind", "category": "code", "one_liner": "Converts CSS declarations to Tailwind CSS utility classes. Maps standard CSS properties to Tailwind utilities including colors, spacing, typography, layout, borders, and more. Outputs space-separated class strings or @apply blocks.", "replaces": "Manual CSS refactors and single-purpose style converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "css/css-to-tailwind.js", "tests": "1/1", "from": "css", "to": "tailwind", "public": false }, { "slug": "cc-csv-to-json", "name": "csv-to-json", "category": "code", "one_liner": "Convert RFC 4180-style CSV text to formatted JSON through the CLI, browser workbench, and bounded local MCP. Preserve quoted fields and empty cells, report repaired rows, protect special object keys, and expose no arbitrary paths or flags.", "replaces": "Ad-hoc format massaging in spreadsheets and converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "data/csv-to-json.js", "tests": "6/6", "from": "csv", "to": "json", "public": false }, { "slug": "cc-csv-to-sql", "name": "csv-to-sql", "category": "code", "one_liner": "Convert CSV data to SQL INSERT statements. Supports CREATE TABLE generation, type inference, batch inserts, and proper SQL escaping.", "replaces": "Ad-hoc format massaging in spreadsheets and converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "data/csv-to-sql.js", "tests": "1/1", "from": "csv", "to": "sql", "public": false }, { "slug": "cc-csv-to-typescript", "name": "csv-to-typescript", "category": "code", "one_liner": "Convert CSV data to a TypeScript interface and typed const array. Infers types from data values and generates clean, exportable TypeScript code.", "replaces": "Ad-hoc format massaging in spreadsheets and converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "data/csv-to-typescript.js", "tests": "1/1", "from": "csv", "to": "typescript", "public": false }, { "slug": "cc-curl-to-axios", "name": "curl-to-axios", "category": "code", "one_liner": "Convert cURL commands to Axios JavaScript code. Produces a ready-to-run script using axios with proper config for headers, body, auth, params, form data, file uploads, and redirect handling.", "replaces": "Hand-translating cURL snippets into client code", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "api/curl-to-axios.js", "tests": "1/1", "from": "curl", "to": "axios", "public": false }, { "slug": "cc-curl-to-fetch", "name": "curl-to-fetch", "category": "code", "one_liner": "Convert cURL commands to JavaScript fetch() code. Handles headers, body data, JSON, form data, basic auth, cookies, file uploads, and insecure/redirect flags.", "replaces": "Hand-translating cURL snippets into client code", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "api/curl-to-fetch.js", "tests": "1/1", "from": "curl", "to": "fetch", "public": false }, { "slug": "cc-curl-to-go", "name": "curl-to-go", "category": "code", "one_liner": "Convert cURL commands to Go net/http code. Generates idiomatic Go with http.NewRequest, proper header setting, strings.NewReader or bytes.NewBuffer for body, json.Marshal for JSON payloads, and the standard if err != nil error handling pattern.", "replaces": "Hand-translating cURL snippets into client code", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "api/curl-to-go.js", "tests": "1/1", "from": "curl", "to": "go", "public": false }, { "slug": "cc-curl-to-php", "name": "curl-to-php", "category": "code", "one_liner": "Convert cURL commands to PHP code. By default generates native PHP cURL (curl_init/curl_setopt). Pass --guzzle to output Guzzle HTTP client code. Handles headers, body, JSON, form data, file uploads, basic auth, cookies, and SSL verification.", "replaces": "Hand-translating cURL snippets into client code", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "api/curl-to-php.js", "tests": "1/1", "from": "curl", "to": "php", "public": false }, { "slug": "cc-curl-to-python", "name": "curl-to-python", "category": "code", "one_liner": "Convert cURL commands to Python requests code. Generates idiomatic Python using the requests library with proper dicts for headers, json= for JSON payloads, data= for form data, files= for uploads, and auth tuples for basic authentication.", "replaces": "Hand-translating cURL snippets into client code", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "api/curl-to-python.js", "tests": "1/1", "from": "curl", "to": "python", "public": false }, { "slug": "cc-env-to-typed-config", "name": "env-to-typed-config", "category": "code", "one_liner": "Convert .env file to TypeScript config loader with validation. Generates a TypeScript interface, parse function, and validation with descriptive errors. Supports type coercion: PORT=number, ENABLED=boolean, DATABASE_URL=string. Includes Zod-style validation pattern using pure JS.", "replaces": "Ad-hoc format massaging in spreadsheets and converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "data/env-to-typed-config.js", "tests": "1/1", "from": "env", "to": "typed-config", "public": false }, { "slug": "cc-esm-to-commonjs", "name": "esm-to-commonjs", "category": "code", "one_liner": "Converts ESM modules to CommonJS: import → require(), export → module.exports, default export → module.exports =, named exports → exports.x, import.meta.url → __filename.", "replaces": "Manual require/import rewrites and copy-paste markdown conversion", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "module/esm-to-commonjs.js", "tests": "1/1", "from": "esm", "to": "commonjs", "public": false }, { "slug": "cc-eval-remover", "name": "eval-remover", "category": "code", "one_liner": "Finds and refactors eval() and new Function() calls to safer alternatives. Converts eval(\"...\") to JSON.parse where applicable, new Function() to inline arrow functions, and setTimeout/setInterval with strings to function wrappers. Adds warnings for cases requiring manual review.", "replaces": "Ad-hoc sanitizing regexes and unsafe copy-paste fixes", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "security/eval-remover.js", "tests": "1/1", "from": null, "to": null, "public": false }, { "slug": "cc-hardcoded-colors-to-tokens", "name": "hardcoded-colors-to-tokens", "category": "code", "one_liner": "Replaces hardcoded color values (hex, rgb, rgba, hsl, hsla) with design token references as CSS custom properties. Generates companion tokens file. Preserves alpha channels via rgba/hsla with token references.", "replaces": "Tool migration guides followed by hand, file by file", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "config/hardcoded-colors-to-tokens.js", "tests": "1/1", "from": "hardcoded-colors", "to": "tokens", "public": false }, { "slug": "cc-html-to-jsx", "name": "html-to-jsx", "category": "code", "one_liner": "Converts HTML to JSX. Handles attributes (class→className, for→htmlFor, style strings→objects, event handlers→camelCase), self-closing tags, boolean attributes, HTML comments, and inline style camelCasing.", "replaces": "Framework rewrite chores and one-off online HTML/JSX converter sites", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "core/html-to-jsx.js", "tests": "1/1", "from": "html", "to": "jsx", "public": false }, { "slug": "cc-html-to-markdown", "name": "html-to-markdown", "category": "code", "one_liner": "Converts HTML to Markdown. Handles headings, paragraphs, links, images, lists, code blocks, blockquotes, bold, italic, tables, and more. Strips unsupported attributes with warnings.", "replaces": "Manual require/import rewrites and copy-paste markdown conversion", "primitive": "Deterministic single-purpose text transform: a zero-dependency Node script, stdin/stdout pipeable, that suggests the right sibling tool on mismatched input", "status": "done", "runner": "converter", "script": "module/html-to-markdown.js", "tests": "1/1", "from": "html", "to": "markdown", "public": false }, { "slug": "cc-html-to-svelte", "name": "html-to-svelte", "category": "code", "one_liner": "Converts HTML to a Svelte component. Maps event handlers to Svelte's on:click syntax, adds