Adobe Firefly Adds AI Music, Speech and Sound Effects Tools

Adobe has rolled out three new audio tools inside Firefly: Generate Music, Generate Speech, and Generate Sound Effects, all now generally available. The update also introduces a free tier for the Firefly AI Assistant with daily generations, adds Google Gemini Omni Flash to the model lineup, and brings new workflow skills like Create Storyboard and Create Brand Kit. The changes target video creators, marketers, and brand teams who spend significant time hunting for licensed audio before publishing.

What happened

Feature	Detail
Generate Music	Powered by Firefly Music Model; creates original tracks matched to video length and mood; universally licensed for commercial use
Generate Speech	Powered by Firefly Speech Model; converts scripts to voiceover with voice, pacing, and emotion controls; ElevenLabs available as an alternative
Generate Sound Effects	Powered by Firefly Audio Model; creates custom sounds based on action, timing, and content energy
Firefly AI Assistant free tier	Daily generations now included at no cost; new skills: Create Storyboard and Create Brand Kit
Gemini Omni Flash	Newly added to Firefly; supports video, audio, image, and text inputs
Berklee study stat	80% of surveyed creators, musicians, and marketers post video daily or several times a week

Adobe has made all three audio tools generally available inside Firefly, the platform that already combines Adobe’s own image, video, and design tools with third-party models from Google, ElevenLabs, Kling AI, Luma AI, OpenAI, and Runway. The audio additions are the most significant expansion of Firefly’s output types since video support landed.

Generate Music: licensed tracks in seconds

The Generate Music tool builds original tracks from a video’s length and mood. Adobe states the output is universally licensed and safe for commercial use, which matters a great deal for anyone delivering work to brand clients. Creator Esther Rehema, quoted by Adobe, said the process that used to take hours now takes seconds. Madeline Salazar, who produces short-form video for brands, said the tool addresses music licensing friction when handing off projects to clients.

Generate Speech and Sound Effects

Generate Speech takes a written script and converts it into a voiceover. Users control voice, pacing, and emotion, and can swap Adobe’s Firefly Speech Model for ElevenLabs if they prefer that voice quality. Generate Sound Effects builds custom audio based on the action and timing described for a scene, pulling from the Firefly Audio Model rather than a library of pre-recorded clips.

Firefly AI Assistant gets a free tier and new skills

The Firefly AI Assistant, which launched in beta earlier this year, now has a free experience with a daily generation allowance. According to Adobe, Create Storyboard and Create Brand Kit are among the most-used skills already. The assistant can also produce branded mockups and batch-edit dozens of images in one pass.

Gemini Omni Flash joins the model roster

Google Gemini Omni Flash is now available inside Firefly alongside the existing lineup. It accepts video, audio, image, and text as inputs, which Adobe says can help convert a rough idea into a storyboarded first cut without switching tools. This continues Adobe’s pattern of pulling in external models rather than building every capability in-house.

Why it matters

For content teams and agencies, the commercial licensing on Generate Music removes a genuine bottleneck. Royalty-free music libraries are large but finding the right track still takes time, and anything outside a cleared library can trigger copyright claims on platforms like YouTube or TikTok. An AI tool that generates a track matched to the edit, with a clean license, cuts that risk and that time.

The ElevenLabs integration inside Generate Speech is worth noting. Adobe is not forcing you into its own voice model, which suggests it is more interested in keeping creators inside the Firefly ecosystem than in pushing a single proprietary pipeline. The same logic applies to the growing model roster (OpenAI, Runway, Luma, Kling, Google), where Firefly acts as a front-end hub rather than a single model.

A Berklee Emerging Artistic Technology Lab study, supported but not controlled by Adobe, found that every respondent uses music in their videos, and 80% post video content daily or several times a week. That is the market Adobe is pitching to.

Our take

The commercial licensing story is the strongest part of this release. Most AI audio tools still leave creators in murky legal territory, particularly for client work. If Adobe’s “universally licensed” claim holds up in practice, Generate Music becomes a useful default for brand content workflows, not just personal projects.

The free Firefly AI Assistant tier is a smart acquisition move. Once teams start building storyboards and brand kits inside Firefly for free, upgrading feels like a natural next step rather than a cold purchase decision. That said, the real test is whether the audio quality is good enough to ship to clients without embarrassment.

If you are already using tools like AI integration workflows that pipe content through multiple platforms, Firefly’s expanding model roster (now including Gemini Omni Flash) could reduce the number of tools you juggle. That is worth a hands-on test before committing. And if audio licensing has been a friction point in your video projects, track how much time Generate Music actually saves versus a well-curated library like Epidemic Sound, where you already know the catalog.

For more context on how AI tools are reshaping creative workflows, see our coverage of why public trust in AI products is still fragile, which shapes how clients respond when you tell them AI made their soundtrack.

What to do about it

1. Open Firefly and test Generate Music on a recent video project to compare the licensed output against your current library workflow.
2. Check Generate Speech against your existing voiceover process; toggle between the Firefly Speech Model and ElevenLabs to find the voice quality that fits your brand.
3. If you run a content team, activate the free Firefly AI Assistant tier and trial Create Storyboard on an upcoming project before deciding whether a paid plan is justified.
4. Review any client contracts that specify music sourcing requirements to confirm Adobe’s universal license covers your delivery obligations.

Test the audio quality on a real client project first. The license terms are only valuable if the output sounds good enough to use.

Source: Bing News · Runway (AI video)