Post by DeepInfra

We just added text-to-music model to DeepInfra. ACE-Step v1.5 is an open-source music foundation model that generates complete songs from a text prompt — vocals, lyrics, and full instrumentation. Not just background music or short clips. Full songs, at quality that rivals commercial tools. A few things worth knowing about how we're running it: → We serve the XL checkpoint, which is the highest-quality variant → The planning step ("thinking") is on by default — it prioritizes musical structure and coherence over raw generation speed → Pricing is $0.001 per second of audio If you're building apps that need music generation — game soundtracks, content creation tools, social video, anything audio — this is now a one-API-call away. → https://lnkd.in/gv9uQ2-r