wtf is savevidai?

SaveVid AI is a tool for downloading videos from Twitter and X. You paste a link to a post, see a preview of the video with details like the author and a thumbnail, choose a video quality, and download the file. The project was built specifically as a reaction to other downloader sites that are full of popup ads, fake download buttons, and redirects. It works without any of that, and it does not track visitors. Under the hood, the tool looks up tweet details using a public service called FixTweet, then streams the actual video file through its own server, since Twitter blocks browsers from reading video files directly from other websites. This is what lets the site show a real progress bar while downloading and give the file a clean name based on the account handle, tweet ID, and resolution. If FixTweet ever goes down or changes its data format, the project has a backup service and a test script to check whether downloads still work. You can run your own copy of this tool with a single Docker command, or deploy it using a provided configuration file for the Render hosting platform, or on any server using Docker Compose and Caddy. The project has both a backend, built with Python, and a frontend, built with TypeScript, each with its own setup steps and test commands for anyone who wants to contribute or modify it. The hosted version optionally collects basic anonymous usage numbers, such as visit counts and two letter country codes, so the maintainer can track growth and notice when video extraction breaks. This is turned off unless specific configuration values are set, so anyone running their own copy collects nothing by default. There is no cookie tracking and no stored identifier tied to a person, and stored records are removed after 90 days. The project is released under the MIT license, which allows free use, including commercial use, as long as the copyright notice is kept.