New Workflow Expansion
PixelForge LLC has introduced a reference-to-video workflow to Flow AI Video, its browser-based AI video creation service. The new capability, available as of September 5, 2026, allows users to generate short video clips from a reference image, complementing the existing text-to-video and image-to-video workflows. The update adds a third input route for users who work with written prompts and visual reference material.
The reference-to-video feature is integrated directly into the Flow AI Video interface, which operates entirely within major desktop and smartphone browsers, eliminating the need for standalone applications. The same web interface is used to select a model, configure supported output settings, review completed results, and download files.
Technical Specifications
The new workflow supports multiple aspect ratios, including 16:9, 9:16, and Auto, where available, allowing creators to tailor output for different platforms and use cases. Users can select from 720p or 1080p resolutions, with video durations of 4, 6, or 8 seconds, depending on the chosen model and subscription plan. Generated videos can be previewed directly in the browser and downloaded as MP4 files for further editing or distribution.
Credit costs and available settings vary by model and plan. The interface shows supported options for the selected model before a generation is submitted. PixelForge has not disclosed specific performance metrics or made comparative claims about the quality of the generated videos, focusing instead on the functional capabilities of the workflow.
Context and Availability
Flow AI Video previously offered text-to-video and image-to-video generation. The reference-to-video addition changes the available input methods without replacing those existing routes. Users can continue to choose an input method according to the source material available for a project.
The reference-to-video workflow is available through the Flow AI Video web interface where supported by the selected model and plan. Current access details are provided on the Flow AI Video website.
The September 5 release changes the browser interface and its available generation routes; it does not change licensing or ownership terms. Users remain responsible for having permission to use uploaded reference material and for reviewing generated output before publication.
Release Scope
The feature was released as a web-service update rather than a separate application. It applies to new generation requests created after the rollout. Previously generated files and the existing text-to-video and image-to-video entry points remain available according to account history and plan rules. Processing availability can change with model capacity and network conditions, and the platform does not guarantee that every output setting will be present for every request.