Adobe Firefly Adds Music, Speech and Sound Effects, and a Free AI Tier

Adobe made Firefly's audio generation tools generally available on 20 August 2026: music, speech, and sound effects, all inside the same creative AI studio that already handles image and video generation. Alongside it, Adobe launched a free tier for Firefly AI Assistant and added Google's Gemini Omni Flash as a model option for generation.

Bringing audio into the same studio as image and video is the more significant move than any single new capability. Generating a piece of AI music has been possible through standalone tools for a while. Generating it inside the same canvas as the video it will sit under, with the same brief and brand context already loaded, removes an entire export and re-import step that used to break creative flow. A sound designer working inside Firefly can now generate a scratch soundtrack, adjust it against a rough cut, and swap it later, all without leaving the tool the video was built in.

The free AI Assistant tier matters more for adoption than for capability. Most of Adobe's AI tooling has sat behind paid tiers, which keeps casual experimentation low and the learning curve steep for anyone deciding whether to move a workflow onto it. A free entry point for AI Assistant specifically lowers the bar for exactly the kind of exploratory use, working out what a tool actually does with a set of brand assets, that turns into a paid adoption decision later.

Adding a third party model like Gemini Omni Flash as an option inside Firefly is also worth noting. Adobe has generally kept its own Firefly models front and centre. Opening the door to outside models suggests less concern about the model itself being the moat, and more focus on the studio, the brief context, and the brand controls wrapped around it. For a studio like ours, that is the part actually worth paying for: not which model generated an asset, but whether the tool respects the brand rules we set before generation happens, not after.