AudD

About AudD

AudD is a music recognition API that identifies songs from audio files, user-generated content, microphone recordings, and live streams, matching against a database of over 160 million songs using its own neural-network audio fingerprinting technology. Sending audio returns metadata including artist, title, album, label, release date, and identifiers such as ISRC and UPC. Official SDKs are available for Python, Node.js/TypeScript, Go, Rust, PHP, Swift, Kotlin, C#/.NET, Java, C, and C++, along with an MCP server for AI agents. Customers include Sony, Warner Music Group, and Universal Music Group, along with music distributors, labels, PROs, video platforms, and AI startups.

The service is used for content analysis — detecting music in short videos, DJ mixes, and other files for trend analysis, recommendation systems, or usage reporting — and for copyright protection, such as scanning social media content, distributor catalogs, user uploads, or AI training data for copyrighted works. A separate stream monitoring product identifies songs playing on radio stations, TV audio, Twitch, YouTube, and other live streams in real time, delivering results by webhook or long polling for use cases like airplay logs, royalty reporting, radio charts, and now-playing displays; an official Twitch extension and embeddable widget can show the currently playing song.

Pricing for file recognition starts with 300 free requests on signup, then $5 per 1,000 requests pay-as-you-go, dropping to about $2 per 1,000 at volume (100,000 requests/month for $450, 200,000 for $800, 500,000 for $1,800). The standard API endpoint analyzes up to about 12 seconds of audio per request; an enterprise endpoint processes files of any length and returns every track with timestamps. Stream monitoring costs $45 per stream per month against AudD's database, or $25 per stream per month against a customer-uploaded catalog, with on-premise deployment and other enterprise arrangements priced individually.

Key features

• Recognizes music in audio files, videos, and microphone recordings against a 160-million-song database
• Real-time recognition on radio stations and audio/video streams via webhook or long polling
• Returns artist, title, album, label, release date, and identifiers like ISRC and UPC
• Enterprise endpoint processes full-length files of any length with per-track timestamps
• Official SDKs for Python, Node.js, Go, Rust, PHP, Swift, Kotlin, C#, Java, C, C++, plus an MCP server for AI agents

Frequently asked questions