Adobe has released its AI generator for transforming text and images into video for public use. The tool, named Generate Video, is now in public beta, following a limited early access phase conducted last year. Users can access the beta tool through the revamped Firefly web app, which also introduces new image generation, translation functions, and subscription tiers for AI credits tailored for creators. Launched in October, Adobe first introduced tools based on its generative AI Firefly Video Model, beginning with the beta version of the Generative Extend tool for Premiere Pro, aimed at extending video footage at the beginning or end. Today's introduction of the Generate Video tool comes more than two months after OpenAI's Sora was revealed. This new tool incorporates minor enhancements since its preview in September. The Generate Video tool features two main components: Text-to-Video and Image-to-Video. The Text-to-Video feature allows users to create videos from text descriptions, while the Image-to-Video feature lets users include a reference image to inform their video prompts. The tool provides various options for refining the output, including adjustments for style, camera angle, motion, and shooting distance. Videos are produced in 1080p resolution at 24 frames per second, a significant improvement from the original 720p quality. Both Text-to-Video and Image-to-Video processes take 90 seconds or longer to generate clips, which can last up to five seconds—shorter than the 20-second clips offered by Sora. Adobe is also working on a faster, lower-resolution "ideation model" and a future 4K version, both of which are expected to launch soon. The Firefly web app, which hosts many of Adobe's generative AI tools, has received a design overhaul and now integrates seamlessly with Creative Cloud applications like Photoshop, Premiere Pro, and Express, streamlining the movement and editing of AI-generated assets.

Notably, Firefly is trained on public domain and licensed content, making it compliant for commercial use. Adobe markets the Generate Video tool as "production-ready, " appealing to filmmakers looking to incorporate AI-generated content without breaching copyright laws. As Adobe enters a competitive landscape in the AI video sector, it faces challenges from rivals such as Sora, which recently launched, and Google, which is beta-testing its next-generation Veo AI video model—seen as an upgrade over the existing OpenAI offering based on early demos. Companies like ByteDance and Pika Labs have also introduced new generative AI video tools. While Adobe's key strength lies in Firefly's commercial applicability, it must continue to enhance the quality and features available against its competition. Additionally, the Firefly web app now offers two more tools in public beta, although they require payment. Scene to Image allows users to create custom references for AI-generated images using built-in 3D and sketching functionalities—stemming from the earlier "Project Scenic" experiment announced in October. The Translate Audio and Video tool enables users to translate and dub audio into over 20 languages while retaining the original speaker's voice. Adobe is rolling out two new subscription plans for Firefly that provide credits for using its models. The Firefly Standard plan starts at $9. 99 per month, offering 2, 000 video/audio credits for up to 20 five-second 1080p video generations. The more expensive Firefly Pro plan is available starting at $29. 99 for 7, 000 credits, allowing for up to 70 five-second 1080p video generations. A significant benefit of both plans is that they provide unlimited access to Firefly's imaging and vector capabilities.