Adobe just dropped some serious updates to Firefly that will change how you make videos. The biggest news? You can now create custom sound effects just by talking into your mic. Say "thunder clapping" or "glass breaking" and Firefly generates the sound effect with the right timing and intensity. Even better, Adobe is adding AI models from its competitors — including Runway's Gen-4 and Google's Veo3 — right inside the Firefly app.
Key Points:
- Record yourself making sound effects and Firefly matches the timing and energy in generated audio
- Access to Runway, Google Veo3, and other AI video tools without switching apps
- New composition reference, style presets, and keyframe cropping for precise editing
The sound effects feature is genuinely innovative. You can record yourself acting out a sound effect, and Firefly uses your voice to guide the timing, volume, and intensity of the generated audio. This gives creators way more control than just typing "explosion sound" and hoping for the best.
Adobe's strategy is smart — instead of trying to beat every competitor, they're bringing the best tools into one platform. The company recently added Runway’s Gen-4 Video and Google Veo3 with Audio to Firefly Boards, and Veo3 with Audio in Generate Video. They also noted that Topaz’s image and video upscalers and Moonvalley’s Marey will be launching soon in Firefly Boards. This means you can experiment with different AI models and tools without jumping between different apps and workflows.
They also added new video controls that are aimed at professional creators. Composition Reference lets you upload a reference video and generate new content that matches the original's structure. Style Presets apply looks like claymation or anime with one click. Keyframe Cropping lets you upload first and last frames and generate video that fits your exact format.
Adobe's biggest advantage remains its focus on commercial safety. While competitors like OpenAI's Sora face questions about training data, Adobe's Firefly models are trained only on content Adobe has permission to use. For businesses worried about copyright issues, this is huge.
The company is also rolling out Text to Avatar, which turns scripts into avatar-led videos. Choose from diverse avatars, customize backgrounds, and select accents — useful for training videos, social content, or pitching ideas.
Adobe's approach differs from OpenAI's Sora, which can generate longer videos up to 20 seconds. Firefly focuses on five-second clips at higher quality. Adobe's bet is that shorter, more usable clips matter more than longer ones that might be harder to edit.
For video creators, this update positions Firefly as a serious one-stop shop for AI video tools. You get Adobe's commercially safe models plus access to cutting-edge competitors, all with professional-grade controls. The voice-guided sound effects alone could save hours of searching through stock audio libraries.
The AI video space is moving fast, and Adobe is making sure it doesn't get left behind by embracing both its own technology and its competitors' best work.