Vellm is a web-based application that generates AI videos from still photographs, using templates modeled after trends on TikTok and Instagram Reels. The service allows users to transform a pet or portrait photo into a short dance or scene video by selecting a template, uploading the required photo or photos, and receiving a generated video file. Templates are organized for cats, dogs, and people, and each one displays its routine or choreography before the user commits to it. The workflow consists of five steps. Users first choose a template based on the trend they want to recreate. They then upload the one or two photos the template requires, and the credit cost is displayed upfront. Generation typically takes 10 to 15 minutes, and the user receives a download link by email. If generation fails, the credits are returned to the user's account. Videos are exported at 1080p resolution and without a watermark. Most templates offer a silent version for posting and an optional version with the original clip's sound for personal viewing. A public-use music license is not included. Generation uses paid credits. See https://vellm.app/pricing for current prices. A one-time purchase can be made without creating an account.