Adobe Firefly Launches New Feature for Multilingual AI Voiceovers

At its annual Adobe MAX 2025 conference on October 28, Adobe announced new generative-AI capabilities for its Firefly creative platform.

Users can now turn text into voiceovers with a new feature called Generate Speech, extending Firefly beyond images and video generation.

According to Adobe’s product announcement, Generate Speech lets users create natural-sounding voiceovers in more than 20 languages and a range of accents, with controls for tone, pacing, emotion, and pronunciation. The company says there are over 70 voices available, developed both by Adobe and “trusted partners.”

Users can choose between Firefly’s foundation speech model — trained on “licensed, commercially-safe” data — or a partner model, such as ElevenLabs Multilingual v2.

“We are integrating great models from across the industry so you have the choices that you need,” David Wadhwani, President of Digital Media at Adobe, said during the keynote.

The workflow is simple: users paste or upload a script, select a voice and accent, adjust settings such as tone, pace, or pronunciation, preview the result, and download the final audio.

Adobe notes that text-to-speech is commonly used to create voiceovers for videos, podcasts, and marketing content, and is also ideal for e-learning and tutorials that need clear narration. 

“Firefly’s AI voice generator makes it easy to add natural, high-quality voice narration to anything from podcasts to training content — improving accessibility and inclusivity for users,” the company said.

The Firefly suite also includes AI translation and AI dubbing features that let users translate audio files and dub videos into up to five target languages per file. “Whether you’re localizing podcasts, voiceovers, or training content, Firefly makes high-quality audio translation easy,” Adobe said. AI dubbing also supports lip sync, helping speakers look and sound natural in the target language.

Adobe says Generate Speech is currently in public beta, but its output is cleared for commercial use.

“Firefly’s AI voice generator makes it easy to add natural, high-quality voice narration to anything from podcasts to training content.” — Adobe

Alongside Generate Speech, Adobe introduced Generate Soundtrack, a fully featured AI music generator that creates instrumental background music from text prompts or video analysis. Users can choose from suggested moods, genres, energy levels, and tempos, or give specific creative directions. The feature is designed for story-driven, commercially safe, instrumental background music — not vocals or songwriting — and supports tracks from five seconds to five minutes. Adobe says all tracks are “commercially safe and completely royalty-free.” Generate Soundtrack is currently in public beta as well.

Generate Speech and Generate Soundtrack are premium features. Free users can try each twice, while continued access requires a Firefly Standard or Firefly Pro plan.

Together, the new features position Adobe Firefly as an end-to-end environment for creative and multilingual media production, covering everything from text and images to voice, music, and localized video.