Mureka

The second-ranked song generator on both music boards — Mureka V9 at 1187 instrumental / 1158 vocals, within 2 Elo of suno V5.5 on instrumental, with V8 third on both (music-arena-leaderboard, 2026-08-03). Absent from this hub until that board was added; its own documentation was read the same day.

The API surface, which is the surprise

Broader than any competitor this wiki holds. From the official platform docs: song generation from lyrics or prompt, instrumental generation, lyrics generation and extension, vocal cloning, song extension, song recognition and description, music transcription, soundtrack generation, remixing, stem separation, file upload — and, unexpectedly for a music company, text-to-speech and podcast creation.

Two things follow. First, Mureka is not only a generator: recognition, transcription and description are analysis, which puts it partly in the same territory as audio-flamingo-3‘s audio comprehension rather than purely in audio-music-generation. Second, the TTS and podcast endpoints cross into this spoke’s text-to-speech branch, making Mureka the second provider after elevenlabs to span speech and music — evidence for the synthesis‘s convergence read.

The vocal cloning endpoint deserves its own note: it is offered on the Pro tier with no consent language anywhere in the material read, the same silence free-voice-clone-list showed across the open catalogue and audio-deepfake tracks as the supply-side gap.

Models, and a version the docs don’t mention

The docs name Mureka V7.5 (“improved melody and arrangement quality… more realistic vocals and instrumental sounds”, streaming playback) and Mureka O1 (curation over generated takes, “multilingual compositions, contextual BGM creation, and cutting-edge AI music editing”).

Neither is V9. The board ranks V9 and V8; the platform documentation still describes V7.5 and O1. So the docs lag the shipped model by at least two versions, and this wiki has a benchmark for a model whose own vendor hasn’t documented it. Recorded rather than reconciled.

The O1 naming — a curation layer that selects among generated candidates rather than a bigger generator — is a sampling-and-selection move, and worth watching if it recurs.

Pricing

Free tier, then Basic $8/month (400 songs, MP3, commercial licensing) and Pro $24/month (1,600 songs, MP3/WAV/stems, advanced editing, voice cloning). That undercuts suno (Pro $10 / Premier $30) and udio ($10 / $30) at the top tier while offering more monthly generations.

Rights: asserted, not evidenced — a third position on the axis

The marketing site states that “every piece of music comes with full commercial rights and usage rights for unlimited applications,” describes output as royalty-free, and names Spotify, TikTok and YouTube as safe destinations.

What is not disclosed anywhere in the pages read: training-data sources or their licensing, any rights-holder agreements, how ownership of output transfers, or any limitation on that blanket commercial grant. Nor is the operating company named on its own site or docs — a search result attributes Mureka to Skywork AI (Kunlun Tech’s brand), which this wiki records as unverified because no first-party page states it.

That places Mureka at a third point on ai-music-copyright‘s axis. suno is top quality under active litigation; udio bought its position with UMG, Warner, Merlin and Kobalt deals; stable-audio trained on a licensed dataset. Mureka is near-top quality with the rights question simply asserted — neither litigated nor licensed, so far as any first-party source here says. An output-side promise (“you may use this commercially”) is not an input-side claim (“we were entitled to train on this”), and only the second is what the lawsuits are about.

Tier

T3 — first-party documentation and marketing. Fine for the API surface, the model names and the pricing, which are the vendor’s own facts to state. The rights claim is a vendor assertion with no evidence attached, and the Elo scores it is paged for come from the neutral music-arena-leaderboard (T2), not from here.

music-arena-leaderboard · suno · udio · stable-audio · ai-music-copyright · audio-music-generation · voice-cloning · audio-deepfake · text-to-speech · elevenlabs · synthesis