Every songwriter's catalog management system starts the same way and breaks the same way. Folders by year, then folders by project, then a spreadsheet, then the spreadsheet stops getting updated somewhere around song four hundred. The system does not fail loudly. It fails when a brief closes and you realise afterwards that you had the right song the whole time.
This page covers what music catalog management software actually has to do for a working songwriter, the four approaches people use, and where each one breaks down.
Strip away the feature lists and there are five jobs:
Job two is where every manual system dies. Nobody hand-enters BPM and key for nine hundred songs, and nobody keeps doing it for the new ones.
Free, portable, and it works up to roughly a hundred songs or however many you can hold in your head. Beyond that you are searching a filesystem for something you have to already remember. Finder cannot answer "what do I have around 120 BPM that sounds hopeful."
Breaks when: your catalog exceeds your memory, which is a smaller number than most writers expect.
The most common serious system, and a real step up: you can filter, sort and share it. The problem is upkeep. Every column is a manual entry, tagging is subjective and drifts over time — the "energetic" you meant in 2023 is not the one you mean now — and the sheet goes stale within a month of the last time you forced yourself to update it.
Breaks when: maintenance cost exceeds the discipline you have available, usually within a few months.
Disco, Songspace, SourceAudio. Cloud storage with private sharing links, access controls, listener analytics and team permissions. Roughly $25 to $99 a month. Genuinely the right tool once other people need to hear your catalog on their own time.
Tagging on these platforms is still largely something you do, and search is still largely keyword search against the tags you entered. You get organisation and delivery. You do not get a system that knows what your songs sound like.
Breaks when: nothing, exactly — but it does not solve the matching problem, and it stops the day you stop paying.
Pitchkit. Instead of asking you to describe your songs, the software listens to them. Tempo, musical key, energy, spectral brightness and danceability are derived from the waveform, lyrics are optionally transcribed with Whisper, and existing ID3 tags are read where present. A thousand-song catalog indexes in roughly an hour, unattended, once.
Then search stops being keyword search. You paste a brief and the catalog is ranked by fit against tempo, mood, theme and artist reference together.
Breaks when: you need other people to access the catalog, need it on Windows, or need team features. It is single-user local software and does not host anything.
| Folders | Spreadsheet | Disco / Songspace | Pitchkit | |
|---|---|---|---|---|
| Cost | Free | Free | $25–99/mo | $49 once |
| Automatic tagging | No | No | Partial | Yes |
| Lyric transcription | No | No | No | Yes |
| Search by description | No | Keyword | Keyword | Yes |
| Rank against a brief | No | No | No | Yes |
| Filters alternate versions | No | Manual | Manual | Yes |
| Share with supervisors | No | No | Yes | No |
| Team access | No | Sort of | Yes | No |
| Keeps working unpaid | Yes | Yes | No | Yes |
| Windows | Yes | Yes | Yes | No |
TV mix, instrumental, acapella — lets software filter them out later.It is software that keeps an index of every song you have written, with enough information about each one that you can find the right song by describing it rather than by remembering its name. In practice it splits into two kinds: hosted platforms such as Disco, Songspace and SourceAudio, which store your masters in the cloud and let you share private links with music supervisors, and local tools such as Pitchkit, which analyse the audio on your own machine and make the catalog searchable by musical characteristics.
The distinction matters because the two solve different problems. Hosted platforms solve delivery. Local analysis tools solve retrieval.
Start by consolidating everything into a single parent folder, because no tool can index files spread across three drives and a Dropbox account. Then decide which file is the primary version of each song and move alternates — TV mixes, instrumentals, acapellas, cut-downs — into consistently named subfolders, so software can filter them out of results later. Name the primary files as artist and title.
After that, index automatically rather than manually. At a thousand songs, hand-tagging BPM, key and mood is roughly forty hours of work that goes stale as soon as you write more songs. Audio analysis derives those attributes from the waveform in about an hour, unattended, and re-runs on new files as you add them.
Audio analysis can derive tempo, musical key, energy, spectral brightness and danceability directly from the waveform with no input from you, and Whisper can transcribe lyrics locally for theme matching. Pitchkit does both, and reads any ID3 tags you already have on top.
What automatic tagging is genuinely good at is the objective layer: tempo, key, and broad energy. What it approximates rather than knows is the subjective layer — whether a song reads as hopeful or resigned depends on context a waveform does not contain. In practice the objective layer plus transcribed lyrics is enough to narrow nine hundred songs to fifteen candidates, which is the part that was costing you hours.
It is good enough while you are still willing to maintain it, and that is the whole problem. A spreadsheet cannot derive anything on its own, so every row is manual entry, and the tagging vocabulary drifts as your taste changes. Most writers describe the same arc: build the sheet in a burst of organisation, keep it current for a few months, then quietly stop.
The deeper limit is that a spreadsheet does keyword filtering, not ranking. You can filter to 'pop' and 'upbeat'. You cannot paste a brief and get your catalog sorted by how well each song fits it, which is the actual job when a brief has a deadline.
Hosted platforms do not in any meaningful sense — the catalog lives on their servers, so no connection means no catalog. Pitchkit does, because it was built that way: indexing, brief matching and lyric transcription all run on your Mac, and after installation the only thing that ever leaves the machine is an artist name, and only if you have added your own API key for artist-reference lookups.
For writers working on unreleased material or under NDA, the fact that audio is never uploaded is often the deciding factor rather than a nice-to-have.
Many working writers do, and they are not redundant. A hosted platform is what you send other people: a private link a supervisor can open on their phone, with access you control. A matching tool is what you use before that: deciding which five songs out of nine hundred belong in the link.
If you are early and nobody has asked for a link yet, the matching problem usually comes first, and paying a monthly subscription for hosting nobody has requested is a common way to spend money without fixing anything. If supervisors are already asking, get the hosting.
Mac app. Indexes your catalog, matches any brief, builds the pitch folder. Runs entirely on your machine. 14-day money-back guarantee.
Get Pitchkit →