Local music and sound-effect asset manager for Mac and Windows: AI tagging, batch organizing, safe writeback
PureLab is for music collectors, DJs, producers, and sound designers. It scans your local music library and asset library, analyzes BPM, key, genre, mood, instruments, and true source quality on your machine, then hands the results back as drafts for you to confirm before writing them into the files. Batch renaming, CJK text conversion, artwork and lyrics embedding, BWF metadata, UCS classification, and Excel round-trips all happen in the same workbench. Every writeback is logged in history: roll back a single file or restore the whole library from a snapshot. New in 6.0: AI Assistant lets you issue organizing commands in plain language with your own model key.
The music library and the sound-effect asset library are managed separately, each with its own views and columns. Mount multiple disks and folders; the virtualized table scrolls, multi-selects, and edits smoothly with hundreds of thousands of rows.
Computes BPM, key, and Camelot code; AI tags genre, sub-style, mood, and instruments. Take the top guess or the top three. Analysis depth can be quick sampling or segmented full-track analysis.
Uses the spectrum and high-frequency cutoff to classify files as true Hi-Res, true lossless, suspected upsample, or suspected lossy transcode, with a 0–100 score cross-checked against waveform and spectrum views.
Analysis results and manual edits all become drafts first; you compare field-by-field diffs before writing back. Writeback goes to a temp file, swaps atomically, then reads back to verify, and is recorded in history. Roll back a single file anytime; Plus / Pro add multi-file batch rollback and whole-library snapshots.
Batch renaming, tag extraction from filenames or paths, case conversion, find-and-replace; Simplified-Traditional conversion and pinyin for Chinese, romaji for Japanese, romanization for Korean; mojibake repair, artist-name normalization; generating nested folders from tags.
Automatically finds same-folder images to use as artwork; oversized ones are shrunk first. Same-name .lrc / .txt lyrics are written into files in batch.
Writes analyzed tags into the fields your target software actually reads, following that app's rules. Built-in templates include Rekordbox comment, Apple Music, Roon, and generic DJ; comment fields can be updated in segments with conflict detection.
Describe the job in plain language inside the workbench, like "normalize this batch's genres to English". It understands the intent, calls PureLab's tools to build a draft, and writes back only after you confirm. It uses your own OpenAI-compatible endpoint and API key.
Describe what you want, like "an upbeat transition effect", or give it a track and find similar-sounding ones. Three search paths combine keywords, text semantics, and audio features, and they compose with tag and metric filters into advanced queries.
Auto-classify sound effects to the UCS standard, or infer the category from folders and filenames. Pro reads and edits WAV bext and iXML data (Beta), preserving other chunks on writeback. Markers edited in the inspector can be exported as text on Plus / Pro.
A small always-on-top window: search the library, drop chosen tracks in, then drag them into Logic, Ableton Live, and other DAWs. Drag-outs transcode to WAV automatically, at 44.1–96 kHz with 16 / 24 / 32-bit float; global hotkeys supported.
Export tags to XLSX for a colleague or client to edit; on re-import, map the columns, preview the result as a draft, and write back after confirming.
The worst fear in batch tag editing is corrupting a file or making a change you can’t undo. PureLab’s writeback follows a fixed sequence:
There are two snapshot types: a quick snapshot backs up tags only; a full snapshot backs up artwork too. Snapshots and the artwork cache can live on an external drive, and retention count and history days are configurable.
| Analysis | Result | Notes |
|---|---|---|
| Audio math | BPM, key, Camelot code | Pure signal processing, no AI model needed |
| Music tags | Genre, sub-style, mood, instruments | Top1 or Top3; minimum instrument confidence is adjustable |
| Source quality | True Hi-Res / true lossless / suspected upsample / suspected lossy | 0–100 score with the high-frequency cutoff |
| Asset tagging | Mood, genre, sound events, instruments | For sound assets; files over 5 seconds |
| UCS auto-classification | UCS category | Adjustable similarity threshold |
| Fingerprint match | Title, artist, and other metadata | Queries an online fingerprint database (needs the network) |
On writeback you can pick a tag language; genre and mood tags can be mapped to your preferred wording through a translation dictionary. Spelling variants can be unified with “canonical terms”, applied globally or per library.
PureLab is a free download. Plus and Pro are both one-time purchases, and Plus users can pay the difference to upgrade to Pro.
| Capability | Free | Plus | Pro |
|---|---|---|---|
| Local scan, tag editing, drafts, single-file rollback | ✓ | ✓ | ✓ |
| Audio analysis, music tags, fingerprint match | 100 uses each | Unlimited | Unlimited |
| Batch organizing (rename, extract, embed, text ops) | Up to 100 tracks per batch | Unlimited | Unlimited |
| Tag Bridge, source-quality detection | ✓ | ✓ | ✓ |
| Directory-level rename / extract, nested folder generation | ✓ | ✓ | |
| Multi-file batch rollback, whole-library snapshots | ✓ | ✓ | |
| Track marker export | ✓ | ✓ | |
| Asset AI tagging (mood, genre, events, instruments) | 100 uses each | 100 uses each | Unlimited |
| Custom tags, advanced filters | ✓ | ||
| Quick search, meaning search, find similar music | ✓ | ||
| BWF bext / iXML (Beta) | ✓ |
Free quotas are lifetime totals and don’t reset monthly.
Scanning, tag read/write, AI analysis, and search all run on your computer. PureLab does not upload music files, project assets, or search content, and it has no account system. Network is needed in these cases:
Analysis and search run on your machine and don't need the network, but the AI models have to be downloaded once. They work offline after that. Fingerprint matching queries an online database; AI Assistant connects to the model endpoint you configured. PureLab does not upload music files, project assets, or search content.
The free edition can scan, edit tags, and roll back single files. AI analysis has a 100-use lifetime quota per feature, and batch operations cap at 100 tracks per batch. Plus is a one-time purchase with unlimited analysis and batch work, and it unlocks directory-level batch operations, batch rollback, and whole-library snapshots. Pro is also a one-time purchase; on top of Plus it adds unlimited asset AI tagging, custom tags, semantic search, advanced filters, and BWF metadata.
Every writeback is recorded in history, and you can inspect each field's change and roll back a single file. Plus and Pro also support multi-file batch rollback and whole-library snapshots: take a snapshot before a large operation and restore the whole library if something goes wrong.
MP3, FLAC, APE, WAV, M4A / MP4 (AAC, ALAC), OGG, DSF, DFF / DSDIFF. WAV files can also have BWF bext and iXML metadata read and written (Pro, Beta).
Yes, through Tag Bridge. It writes tags into the fields each target app expects, with built-in templates for Rekordbox comment, Apple Music, Roon, and generic DJ. Note: if you batch-rename files or folders, tracks referenced by path in your DJ software will need relinking.
A conversational assistant inside the workbench. Describe the job in plain language and it calls PureLab's filtering and editing tools to build a draft, written back only after you confirm. It uses your own OpenAI-compatible endpoint and API key, stored only on your machine.
DirectML is the default. NVIDIA users can configure CUDA in settings; if DirectML is unstable on an integrated GPU, switch to CPU. The settings include a hardware diagnostic that tests whether the current acceleration path works.
PureLab is designed for offline local processing and does not depend on cloud transcoding. Read ourprivacy policyfor full data-handling details.
You can reach us via email feedback@efficient-lab.com or submit issues and suggestions through GitHub Issues.
Local stem separation for Mac and Windows: split a song into vocals, accompaniment, drums, bass, and more
Local voice denoising and de-reverb for Mac and Windows
Local audio restoration for Mac and Windows: repair recordings with hiss, noise, and missing frequency ranges
Batch-extract a whole folder of SACD ISOs into tagged DSF files