SampleSift 1.0 (MacOS) - From Tweakbench!

[tweakbench]

I'm proud to announce my first pure utitlity application for musicians and DJ's.

SampleSift analyzes your audio library locally — BPM, key, energy, and tags — so you can find the right kick or bass in seconds, not minutes. No cloud, no signup.

SampleSift is the audio sample browser for producers and DJs who have outgrown Finder.

Point it at the folders where your samples live. It scans them, analyzes each file locally for BPM, musical key, energy, and content tags, and gives you a fast, searchable library that stays on your machine — no cloud, no account, no subscription.

WHAT IT DOES
- Audio analysis on your Mac. SampleSift uses Apple's Accelerate and SoundAnalysis frameworks to detect tempo, musical key, energy, and instrument type for every file. Multi-pass: a fast first pass surfaces filename-derived data immediately, with DSP and machine-learning passes refining results in the background while you keep working.
- Smart filters and collections. Narrow your library by BPM range, key, channels (mono/stereo), file format, energy, rating, and tags. Combine filters freely, then save the result as a smart collection you can return to with one click.
- Filename intelligence. SampleSift reads the naming conventions used by sample packs — "kick_120bpm_Cmin.wav", "Lead 140 F#m" — and pre-fills BPM, key, and tags before any audio analysis runs. Your library is browsable in seconds, not hours.
- Audition with waveform preview. Click a file to hear it. Spacebar to play and pause. Loop mode for one-shots-as-loops. Multiple-selection playback for stepping through a folder.
- Export curated sets. Filter to exactly what you want for a session, then Copy To drops them into your DAW's project folder.
- Watch folders. SampleSift sees new files as they land — drop a fresh pack into a watched folder and it appears in your library without a manual rescan.
- Duplicate finder. Find redundant copies of the same audio file across multiple folders by content, not just filename.

PRIVACY
SampleSift collects no data. It does not connect to the network. It does not phone home for telemetry, updates, ads, or anything else. Your sample library — its contents, filenames, tags, and analysis results — never leaves your Mac. The only directories the app can read are the ones you explicitly grant access to via macOS App Sandbox.

REQUIREMENTS
- macOS 14 (Sonoma) or later
- Apple Silicon or Intel
- Works with WAV, AIFF, MP3, M4A, AAC, FLAC, OGG, CAF
- SampleSift is built by Tweakbench, makers of small, focused tools for musicians.

SampleSift 1.0 for MacOS is available now ($4.99 US): https://apps.apple.com/us/app/samplesif ... 8757?mt=12

[Morty-C-137]

Damn.. I have to try this. So many of these sample manager offerings coming onto the market as of late. But a lot of new and unknowns with zero track record in audio plugins business (I know, we all gotta start somewhere).

No phoning home and not needing to be online sounds persuasive.

EDIT: Can we select tags somewhere on the GUI to live filter samples? Say, in the top bar I click on a tag called 'Kicks' and it shows all kick drums?

"Free to try for 7 days, then a one-time $4.99 in-app purchase"

I am not seeing that option anywhere? In the App Store only the buy button requires immediate payment.

[tweakbench]

Sorry, the trial comes with the 2.0 update.. which is stuck in Apple review hell. Any day now!

[tweakbench]

Ok! v2.0 is now live.

[pekbro]

Soooo, no windows version? ever? is that whats going on, due to the lack of apple frameworks? C'mon, you can do it, it's called Audio Feature Extraction.

[tweakbench]

Sorry no, the tech i'm using for ML analysis is specific to Apple's hardware. It's possible one day I could port it to RTX Sparc for windows, but that would require users to have that specific hardware.

[pekbro]

Sorry no, the tech i'm using for ML analysis is specific to Apple's hardware. It's possible one day I could port it to RTX Sparc for windows, but that would require users to have that specific hardware.

I'm just saying that's all they are doing even if they have their own implementation, AFE is a real thing that has nothing to do with apple. But I get it, no worries...

[Morty-C-137]

Ok! v2.0 is now live.

Hmm.. still no 7 day demo on the App Store.

[tweakbench]

lol, sorry. i forgot to set the price to free (but the IAP is active and working). check again now!

[Morty-C-137]

Ok, working now. I will check this out!



Edit: Ok, very lightweight and fast! I can't find any ML resources in the apps install locations, how is it tagging everything so quickly, so fast? (though on second glance I am noticing a LOT of snare drums are getting tagged as kicks).

The only two things I would still like is a preference menu where we can set the audio output (internal or external sound card) as I tend to monitor through my external monitors that get fed by my sound card. And I place in the GUI that displays known tags we can click to quickly sort the library without us having to type in the tags we are looking for in the tag search bar (many implement it the way you have but it always feels clunky).

Same for sorting between loops an one-shots.

[Morty-C-137]

I ran the whole scan again using the ML option and that seems to have done a more thorough job, but to the point of 'over-tagging' ?

A UI issue I am running in to is that the blue and purple colored tags become near to impossible to read once a sample had been selected:



Will keyboard shortcuts be implemented in a future update? Would be helpful to speed things up to have keyboard shortcuts for some of the right click menu stuff.

Overall really loving the premise of this app and will def purchase after the demo period ends

[audio/fault]

downloaded the demo earlier today… still analysing 121k samples around 80% after about 6h

[tweakbench]

downloaded the demo earlier today… still analysing 121k samples around 80% after about 6h

What processor are you running?

[tweakbench]

Thanks for all the feedback, I'll get some revisions in soon and will push an update maybe Monday.

[audio/fault]

downloaded the demo earlier today… still analysing 121k samples around 80% after about 6h

What processor are you running?

Mac Mini M2
It finished a while ago. Will have a closer look on sunday.