An ACRCloud alternative
for the stations you choose

ACRCloud's broadcast monitoring is built around a panel. They index tens of thousands of radio and TV stations continuously, and you query what that panel heard, including years before you signed up. If your question is “where in the world did this record get played”, a panel is the only structure that can answer it, and ours cannot.

Sonovault inverts it. You register the stations that matter to you and we monitor exactly those, continuously, from that moment on. The log belongs to your account rather than being a slice of a shared dataset. That is worse for discovery and better for everything downstream: you know precisely which stations are covered, the price is a function of a number you chose, and the log is yours to query, export, and report from without asking what the panel happened to include.

The honest version of the trade: if you do not know which stations matter yet, a panel is worth paying for. If you do know, paying for a panel is paying for coverage you will not read. That is the case where an ACRCloud alternative is worth looking at, and the case this page is written for.

With a dedicated recognition vendor, the second half is your problem: a second provider, a second contract, and a reconciliation job between two catalogs that disagree about which row is the same recording. That reconciliation is the part teams underestimate.

On Sonovault the match is a key into the same catalog every other endpoint reads. One id, and the ISRC, the ISWC behind the recording, the label, the release date, the genre, and the track's IDs on Spotify, Apple Music, Tidal, Beatport, Discogs, and MusicBrainz are all one call away, on the same key and the same bill.

# Fingerprint locally, send ~1 KB of integers. 10 credits.
curl -X POST https://api.sonovault.now/v1/tracks/identify \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"fingerprint": [1924357, 1924361, ...], "fingerprint_duration": 302}'

# The match is a key into the catalog, not a dead end.
curl "https://api.sonovault.now/v1/tracks/121217540"      -H "x-api-key: KEY"  # ISRC, label, release, genre
curl "https://api.sonovault.now/v1/tracks/links?id=121217540" -H "x-api-key: KEY"  # six-platform IDs
curl "https://api.sonovault.now/v1/tracks/iswc?id=121217540"  -H "x-api-key: KEY"  # the composition behind it

The same applies to a stream. Every row of an airplay report already carries the ISRC, which is the field a royalty workflow needs and the field a play log almost never has. See the stream monitoring reference for the endpoints.

Ours, in full: monitoring is EUR 29 per active station per month, billed only while the station is running, on top of any API plan including the free one. Reading your own now-playing, history, and airplay reports burns no API credits. API plans are free for 1,000 requests a month, then EUR 29 for 50K, EUR 99 for 500K, and EUR 249 for 5M, with a flat 20% off annual billing. Clip recognition costs 10 credits per fingerprint, or 10 plus 1 credit per MB if you upload the audio.

ACRCloud's side of that comparison is the part we cannot write. They do not publish a rate card: their pricing page routes you to console signup or a contact form. So we are not going to print a price comparison, because we would be inventing one half of it, and a comparison built on a guess is worth nothing to you.

What we will say is structural, and you can verify it in thirty seconds. Our price is on a public page and you can multiply it by your station count before you talk to anyone. Theirs requires a conversation first. If you are costing a project this quarter, that difference is the one that actually affects you.

Monitoring vendors generally describe capability and keep the mechanism private. We wrote ours up instead: a dedicated listener per station, overlapping analysis windows, two different recognition methods running side by side with the engine learning which one wins on each station, a per-station hot set held in memory, and a confirmation gate that keeps a track out of your log until several consecutive windows agree.

You can read the whole thing before you spend anything. How the detection engine works covers it in plain language, including where the trade-offs are.