--- title: "APIs to detect copyrighted music in user uploads, compared (2026)" description: "AudD, Audible Magic, Pex (Vobile), ACRCloud, and YouTube Content ID compared for platforms that need to find copyrighted music in user-uploaded video and audio: what each returns, who sets the rights policy, and what scanning costs." slug: "/resources/articles/best-apis-to-detect-copyrighted-music-in-uploads" section: "articles" keywords: [audd, copyright detection api, detect copyrighted music, user-generated content, ugc, audible magic, pex, vobile, content id, music recognition api] --- # APIs to detect copyrighted music in user uploads, compared (2026) A platform that accepts user video or audio needs to know which uploads contain copyrighted music, which recordings they contain, and where. What it does next depends on its licenses: publish, block, mute the segment, or attribute the use to the rights holder. Services split into two kinds. **Identification APIs** tell you what is in the file and leave the decision to you. **Rights intermediaries** also hold rights holders' registrations and policies and return a decision or a license status. ## What to check - **Whole-file scanning.** Uploads are minutes or hours long, and a song can appear anywhere. You need every match and its position in the file. - **Identifiers.** Label, ISRC, and UPC let you look up rights holders and licenses. - **Your own catalog.** Rights holders on your platform, or your own licensed library, may include music missing from commercial catalogs. - **Who sets policy.** Decide whether you want to write the block or monetize rules yourself or receive them from rights holders. - **Cost per hour of audio.** At platform scale this decides between scanning everything and sampling. ## AudD AudD's enterprise endpoint takes a file or URL of any length, including links to pages on YouTube, TikTok, Instagram, SoundCloud, Twitch, Facebook, and X. It returns every recognized track with start and end times in the file, the label, and, on a Startup plan or higher, ISRC, UPC, and a match score. Matching uses AudD's neural-network fingerprinting against 160 million songs, and it recognizes remixes, sped-up and slowed versions, and music under speech. Rights holders and platforms can upload their own tracks to a custom catalog, including unreleased music. Matches against it return the ID you assigned. AudD returns identification only. Your platform applies the policy: block, mute, route to licensing, or allow. The policy logic is yours to build or to get from a licensing partner. **Pricing:** the enterprise endpoint counts one request per 12 seconds of audio scanned, so a one-minute upload is five requests. Requests cost $5 per 1,000 pay-as-you-go and about $2 per 1,000 on volume plans. The `skip` and `every` parameters let you sample long files: `skip=4&every=1` scans 12 seconds out of every minute. Larger volumes are priced on request. **For compliance reviews:** enterprise customers handling sensitive data can sign a data processing agreement. Audio is deleted after processing and is not used to train models. AudD can also be licensed on-premise, so uploads never leave your infrastructure. **Who uses it:** Sony, Warner Music Group, and Universal Music Group use AudD to find their recordings in user uploads. ## Audible Magic Audible Magic positions itself as the intermediary between rights holders and social platforms. Its site names Twitch, SoundCloud, and Dailymotion among its platform customers. Beyond identification, its Rights360 platform determines license status and handles royalty reporting and payment administration. It reports a catalog of more than 100 million songs from over 300,000 labels, publishers, and collecting societies. Audible Magic is sold through enterprise agreements, with prices on request. ## Pex (Vobile) Vobile acquired Pex in April 2025. Pex's Attribution Engine combines an asset registry, where rights holders register works for free and attach a policy (monetize, block, direct deal, or free license), with content identification. A platform-side SDK sends fingerprint hashes of each upload, and the engine returns the rights holders for both recording and composition and the policy for each match. Pex matches clips as short as one second and covers melodies, remixes, and live versions. It also runs a dispute process between uploaders and rights holders. Prices for platforms are on request. Pex's cross-platform discovery product starts at $1 per file per month. ## ACRCloud ACRCloud lists copyright compliance among its products, alongside music recognition, custom databases, and AI-music detection. It is an identification service: like AudD, it leaves policy to the platform. It offers a 14-day trial, with prices on request. ## YouTube Content ID Content ID only works on videos uploaded to YouTube, so it can't protect another platform. YouTube grants Content ID only to rights holders it approves, typically large labels and distributors. ## Matching the service to your platform | You need | Consider | |---|---| | Every track in each upload, with positions, public pricing, and your own rules | AudD | | Scanning against rights holders' own or unreleased catalogs | AudD custom catalog; Pex asset registry | | Rights holders' policies and license status returned with each match | Audible Magic, Pex | | Royalty reporting and payments handled for you | Audible Magic | | Uploads that never leave your servers | AudD on-premise | The two kinds can be combined: an identification API scans every upload, and a licensing partner covers the catalogs it has deals for. ## Trying AudD on your uploads Get a token at [dashboard.audd.io](https://dashboard.audd.io). The first 300 requests are free. Always set `limit` while testing the enterprise endpoint so one long file can't use more requests than you intended. Full reference: [enterprise documentation](https://docs.audd.io/enterprise). For custom catalogs, DPAs, or on-premise, write to [api@audd.io](mailto:api@audd.io). --- **Related** - [How to detect copyrighted music in user-generated content](/resources/articles/detect-copyrighted-music-in-user-generated-content) - [Copyright detection in user-generated content: an implementation guide](/resources/articles/copyright-detection-in-user-generated-content) - [AudD vs Pex](/resources/articles/audd-vs-pex) - [GDPR and music recognition](/resources/articles/gdpr-music-recognition)