Upload a reference audio file, then select a folder to search for similar files using HPCP harmonic fingerprinting (Essentia.js).
Choose a single audio file to use as reference.
Select a folder containing audio files to search through.
Upload one reference audio file, scan a folder, and Sakarto finds everything that sounds harmonically similar to it using Essentia.js HPCP fingerprinting. The reference card is always pinned first — protected from move or delete — and matches appear once the full scan completes.
Reverse Audio Search is the audio equivalent of reverse image search. Instead of scanning a folder and automatically grouping everything that sounds alike, you choose one specific audio file as your reference, then Sakarto scans a folder to find everything with a similar harmonic fingerprint.
Use it when you already know what track you're looking for — for example: "find all copies, covers, and re-encoded versions of this specific song in my archive."
Upload a single audio file. Sakarto processes it immediately and displays a player so you can verify you picked the right file.
Select a folder and Sakarto fingerprints every audio file in it using Essentia.js HPCP, then compares each against your reference using cosine similarity.
HPCP extraction continues even when you switch browser tabs. Unlike video scanning, the Web Audio API is never throttled in background. Essentia.js WASM initialises once and is reused for every file.
The reference card is pinned first with a blue REFERENCE badge. It can never be accidentally moved or deleted.
Both the duplicate finder and this reverse search use the same Essentia.js HPCP fingerprinting pipeline. The difference is what gets compared to what: here, every folder file is compared against your reference only — not against each other.
When you select your reference audio, Sakarto first initialises the Essentia.js WASM module (1–3 seconds on first use), then immediately processes the reference file — decodes it at 44,100 Hz and computes its full HPCP sequence. A preview player appears so you can confirm the file loaded correctly before scanning begins.
For each audio file in the folder, the browser decodes it at the native 44,100 Hz sample rate using AudioContext.decodeAudioData(). Only the left channel is used. Files over 100 MB are skipped. The reference file itself is skipped if it also appears inside the scanned folder (matched by name + size + date).
Each folder file's audio is divided into overlapping 4,096-sample frames (hop 2,048). Essentia's hpcpExtractor(frame, 44100, false) produces a 12-element Harmonic Pitch Class Profile per frame — one bin per semitone of the chromatic scale, capturing the relative harmonic energy at each pitch class.
The shorter of the two HPCP sequences determines the alignment length. Each corresponding pair of 12-element vectors is compared using cosine similarity — measuring the angle between them in 12-dimensional chroma space. These per-frame scores are averaged to produce a single similarity value for the file pair.
A duration pre-check first rejects any folder file shorter than 50% of the reference duration. For the remaining files, if the average cosine similarity meets or exceeds the threshold set by the slider, the file is added to the results list as a match. Files that don't meet the threshold are silently skipped.
Unlike the duplicate finder (which shows groups live), reverse search completes the full folder scan first, then displays all matches together. The reference card is pinned at position 0 with a blue REFERENCE badge. All matching files appear after it. The result count shows how many similar files were found in total.
The Similarity Threshold slider (50–95) controls how strictly folder files must match your reference to appear in results. Higher values are stricter — cosine similarity of 1.0 means a perfect harmonic match. Adjusting it after a scan re-filters results instantly without re-scanning.
Only near-identical files match. Use for exact copies, same file at different bitrates, or lossless vs. lossy exports of the same master. Very few or no false positives.
Catches cover versions, different-bitrate re-encodes, and recordings with the same chord progression. A good starting point for music library searches.
Groups files sharing broad harmonic character — same key, similar chord patterns. Useful for finding harmonically related recordings. Expect more false positives — always listen before deleting.
Every audio card has a built-in player. Here's how to use it:
| Control | What it does |
|---|---|
| ▶️ Play button | Starts playback. Automatically pauses any other playing track on the page. |
| ⏸️ Pause button | Pauses playback. Audio position is remembered. |
| Progress bar | Click anywhere on the bar to seek to that position. Drag left/right to scrub. |
| 🔊 / 🔇 Mute button | Mutes or unmutes this individual track. |
| 🔇 Audio (toolbar) | Mutes or unmutes all audio on the page at once. |
| Time display | Shows current position / total duration (e.g. 1:23 / 3:45). |
| ← / → arrow keys | When the progress bar is focused, seeks backward or forward 5 seconds. |
Click Choose Reference Audio. Select any audio file. Sakarto first initialises the Essentia.js WASM module (1–3 seconds on first use), then processes the reference file immediately — a preview player appears so you can play and confirm you chose correctly.
Supported formats: MP3, WAV, FLAC, OGG, M4A, AAC. Files over 100 MB are not supported as reference files.
Once a reference file is loaded, Choose Folder to Search becomes active. Click it to open a native folder picker (Chrome / Edge). Or drag and drop a folder directly onto the page.
The folder count shows how many files were found. The Start Search button activates only when both a reference file and a folder are selected.
Click 🔍 Start Search. A progress bar shows how many files have been processed. Audio processing runs fully in the background — you can switch browser tabs freely without pausing the scan.
Once the scan completes, the reference card appears at position 0 (blue REFERENCE badge) followed by all matching files sorted by closeness to the reference. The result count shows how many similar files were found.
Use the Similarity Threshold slider above the results to tune sensitivity. Releasing the slider re-filters results instantly without re-scanning. Start at the default of 75 and adjust based on results.
Every card has a built-in player. Use it to listen and confirm similarity before taking action. Playing one track automatically pauses all others.
Ctrl+Click the reference card and one or more match cards, then click ⚖️ Compare. A modal opens with a player per file, full metadata (duration, size, type, path), and a similarity percentage. For 3+ files, a pairwise similarity matrix is shown.
The reference card in the compare modal shows only a Copy Name button — no Move or Delete — to keep it protected.
Select match cards (not the reference — it's protected) and use the toolbar. With Queue Mode on (recommended), files are staged first:
With Queue Mode enabled (default), clicking Move or Delete stages files in the Move Queue or Delete Queue tab. Switch to those tabs in the sidebar, listen to each file, remove any you change your mind about, then execute. The default destination folder is Audio-Duplicates, created inside your scanned folder.
| Button / Control | What it does |
|---|---|
| Choose Reference Audio | Select the audio file to search for. Player preview appears immediately. |
| Choose Folder to Search | Select the folder to scan. Activates once reference is loaded. |
| 🔍 Start Search | Begin scanning. Processes all folder files, then shows all matches. |
| Similarity Threshold slider | Adjusts matching strictness (50–95, higher = stricter). Re-filters results on release without re-scanning. |
| 🔄 New Search | Reload the page and start fresh with a new reference file |
| 📋 Copy (N) | Copy selected filenames to clipboard. N = count. |
| 📂 Move (N) | Move match files or stage them in Move Queue (reference excluded) |
| 🗑️ Delete (N) | Delete match files or stage in Delete Queue (reference excluded) |
| ✕ Deselect All | Remove selection from all cards |
| ⚖️ Compare (N) | Open compare modal for Ctrl+clicked files. Needs 2+. |
| 🔇 Audio | Mute / unmute all audio players on the page |
| Queue Mode checkbox | Stage files for review before executing. Recommended on. |
| Click card | Select / deselect the card |
| Ctrl + Click card | Add / remove from compare list (purple border) |
| Click & drag (empty area) | Box-select multiple match cards |
| Right-click card | Context menu: Copy / Move / Delete / Compare. Move & Delete disabled on reference. |
| 🔍 hover icon | Open full-size preview with modal player. Reference shown as REFERENCE. |
| ▶️ on player | Play audio. Pauses all other players on page automatically. |
Always back up your audio library before deleting anything. Deleted files bypass the recycle bin and cannot be recovered.
Use the highest-quality, longest version of the file as your reference — a lossless FLAC or 320k MP3. A heavily compressed reference produces fewer HPCP frames and less reliable matching.
The default of 75 is well-calibrated for most music. If expected covers aren't showing, lower toward 65. If unrelated songs appear, raise toward 80–85.
Use the built-in players to audition reference and match side-by-side before moving or deleting. The algorithm is good but not infallible for edge cases.
The Essentia.js WASM module loads from CDN on first use. If the progress bar freezes at 0% for more than 15 seconds, check that your network can reach cdn.jsdelivr.net. Some ad-blockers block CDN requests.
If Essentia HPCP finds nothing, lower the threshold to 60–65. For spoken word or sound effects, switch to Meyda MFCC Reverse Search. For exact copy detection without WASM, use Chromaprint Reverse Search.
Essentia.js loads from CDN and compiles once per session. This 1–10 second pause is unavoidable on first use. If it never completes, check network access to cdn.jsdelivr.net. Some firewalls and ad-blockers block CDN requests.
The File System Access API bypasses the OS recycle bin entirely. Use Queue Mode to review files before executing any delete operation.
You can't scan two separate folders in one session. Workaround: scan their common parent directory so both are included.
HPCP stores a full array of 12-element vectors per file (one per frame). For very large libraries, RAM usage is significantly higher than compact hash methods. Scan in subfolders rather than entire drives.
The File System Access API is only available in Chrome 86+ and Edge 86+. Firefox and Safari can scan and display results but cannot move or delete files.
remove() call permanently deletes files without going through the OS Recycle Bin. The reference file is fully protected and cannot be deleted or moved no matter what. For all matched files, always use Queue Mode: stage them in the Delete Queue, play each one in the player, compare it against the reference card in the Compare modal, and only execute the deletion when you're certain.