Adobe Firefly Ships Music, Speech, and Sound Effects to GA

Adobe said on 20 August 2026 that Adobe Firefly’s audio tools are now generally available: Generate Music, Generate Speech, and Generate Sound Effects, with no separate audio subscription. The same post added Gemini Omni Flash and a free Firefly AI Assistant with daily limits.

What Adobe Firefly audio tools now include

Generate Music is powered by the Firefly Music Model. Adobe describes the tracks as commercially licensed originals that can match a video’s length and mood. That is the production claim: a bed that fits the cut, with a license path that is not a scavenger hunt through a stock library.

Generate Speech is powered by Firefly Speech, with ElevenLabs available as an option. The job is script-to-voiceover. If you already keep a house voice in another tool, treat ElevenLabs as an optional speech route, not a second product to buy.

Generate Sound Effects is powered by the Firefly Audio Model. That is Foley-shaped generation sitting next to the music and speech tools, not a second subscription.

Adobe is explicit that this audio stack does not need its own subscription. Music, speech, and effects sit inside Adobe Firefly. Teams that were stitching a music app, a TTS tab, and an SFX pack into one timeline are the audience.

Gemini Omni Flash and a free assistant

The same 20 August update adds Gemini Omni Flash inside Firefly. Prompts can mix video, audio, and image inputs with text. Adobe’s pitch is that a rough idea can become a storyboarded first cut without leaving the studio. That is a model choice, not a replacement for the Firefly-native audio trio.

Firefly AI Assistant now has a free experience with daily generations. Adobe is not publishing a free-tier ceiling in the facts used here, so do not invent one. Daily limits exist. The assistant is free to try. That matters if you write skills: the trigger is an outcome (voiceover, licensed bed, first cut), not a model slogan.

If your library already covers Suno prompts or the Stable Audio 3 DAW plugin beta, this is a third audio surface with a different license story. Suno is still a music generator with its own legal weather. Firefly’s claim is commercially licensed music that matches length and mood, plus speech and effects in the same studio.

How to write the skill so the studio stays one place

Keep the prompt about the cut, not about the brand of model. A useful Firefly audio skill names duration, mood, and whether the output is music, speech, or effects. It then picks Generate Music, Generate Speech, or Generate Sound Effects. If the brief needs a second voice stack, the skill can name ElevenLabs as an optional speech route.

Do not hide Gemini Omni Flash inside a music skill. Multi-modal storyboard work is a different trigger. Video plus audio plus image plus text is a first-cut job. A 30-second licensed bed is a music job. Mixing those in one card is how you get a voiceover when you asked for a sting.

Commercial-use language belongs in the output contract. If a client needs music they can ship, the skill should say “Firefly Music Model, commercially licensed, length-matched” and stop there. Do not upgrade that to a legal opinion. Do not paste a survey number you do not have.

The citation sentence is Adobe Firefly, 20 August 2026, Generate Music, Generate Speech, Generate Sound Effects, generally available, no separate audio subscription. Gemini Omni Flash is in the same post. The free assistant has daily limits. Those are the facts.

Sources

• Adobe Firefly expands its creative AI studio: generate music, speech, and sound effects in one place — Adobe Blog, 20 August 2026