What is the best API for identifying songs from the microphone in a mobile app?

For an Apple-platform app that only needs Shazam's catalog, ShazamKit has no per-request charge. If you need the same code on iOS and Android, recognition from a server, long files, streams, or your own catalog, AudD has official Swift and Kotlin SDKs and costs $5 per 1,000 requests after 300 free. ACRCloud adds humming recognition and offline matching of your own content. AcoustID is built for full, clean files and isn't suited to microphone clips.

Is there a Shazam API?

Apple offers ShazamKit, an SDK for iOS, iPadOS, macOS, tvOS, visionOS, and watchOS, with a separate SDK for Android. There is no general-purpose Shazam REST API for sending audio from a server.

Song recognition APIs for mobile apps compared (2026): ShazamKit, AudD, ACRCloud, AcoustID

Q: How much does song recognition cost for a mobile app with AudD?

300 requests are free on signup. After that it's $5 per 1,000 recognitions pay-as-you-go, or $450 per month for 100,000, $800 for 200,000, and $1,800 for 500,000.

Your app records a few seconds of audio from the microphone, and you want the song’s title and artist back, usually with a link to play it. This article compares four services: Apple’s ShazamKit, AudD, ACRCloud, and AcoustID.

ShazamKit

ShazamKit is Apple’s SDK for matching audio against the Shazam catalog. It runs on iOS 15, iPadOS, macOS 12, tvOS, visionOS, and watchOS, and there is a separate ShazamKit SDK for Android. The SDK turns captured audio into a signature on the device, so the recording itself isn’t sent to Apple. You can also build custom catalogs from your own audio, which is useful for second-screen and sync features.

Access requires an Apple developer account: you enable the ShazamKit service for your app, and on Android you create a media identifier and a private key to sign developer tokens. ShazamKit has no per-request charge.

Limits: ShazamKit is a client SDK. It doesn’t offer a general REST API for sending audio from your server, recognizing whole files, or monitoring live streams. Results come from the Shazam catalog and whatever custom catalog you bundle.

AudD

AudD is a REST API with official SDKs for Swift and Kotlin, among eleven languages. The same recognition call works on iOS, Android, and your backend. The standard endpoint analyzes up to about 12 seconds of audio and responds in under two seconds. It matches against 160 million songs using AudD’s own neural-network fingerprinting.

A match returns artist, title, album, label, release date, and a song_link that opens the song in the listener’s streaming service. Add return=apple_music,spotify for IDs, artwork, and preview URLs from those services. The same API also recognizes long files and live streams and matches custom catalogs.

Pricing: 300 requests free on signup with no card. After that, $5 per 1,000 recognitions pay-as-you-go, or $450 per month for 100,000, $800 for 200,000, and $1,800 for 500,000.

Limits: each request is a network call, so there is no offline mode.

ACRCloud

ACRCloud offers mobile SDKs and a REST API against a catalog it describes as more than 150 million tracks. It also offers humming recognition, which the others lack, and offline recognition of your own content on the device. API requests are signed with HMAC using an access key and secret, and there’s a 14-day trial, with prices on request.

AcoustID

AcoustID is an open-source identification service with a crowdsourced fingerprint database linked to MusicBrainz. It fingerprints whole files with its Chromaprint library and is built for tagging music collections, not for short, noisy microphone clips. It is free for non-commercial use and limited to three requests per second; commercial use requires a paid plan from AcoustID OÜ.

Which one for which app

Your app	Consider
Apple platforms only, Shazam catalog is enough, no server-side recognition	ShazamKit
Same recognition code on iOS, Android, and your server	AudD
Recognition of long files or live streams alongside the app	AudD
Your own catalog matched from the backend	AudD custom catalog
Humming recognition, or matching your own content offline	ACRCloud
Tagging full music files in a library app	AcoustID

Some apps use more than one. TrackRadar, a track-identification tool for DJs, sends audio to AudD first, then Shazam, then ACRCloud when the earlier service finds nothing. Its comparison describes AudD as having the best electronic-music coverage of the three.

Adding AudD to an app

Get a token at dashboard.audd.io, then add the SDK. The SDK documentation has install commands and examples for Swift, Kotlin, and the other languages.

Related