Unshared Projects
Unshared projects on Scratch cannot be opened in Bilup, the packager, or any other third-party site. This is a Scratch restriction, not a Bilup one, and there is no way around it that keeps your account safe.
Any website other than scratch.mit.edu that asks for your Scratch password is a scam, even if it claims to let you open your unshared projects. You will have your account stolen. There are no exceptions to this rule.
Why this happens
Downloading a project from the Scratch API requires a temporary "project token". For unshared projects, that token can only be obtained by the project's owner, and it expires after a few minutes. Even if you are signed in to Scratch in the same browser, Bilup cannot read that token, so it cannot download the project data.
This was a deliberate change by the Scratch Team to make unshared projects actually private. Before it, "unshared" projects were effectively public to anyone who knew the ID, which was a real privacy problem given how many contain personal information. Securing them was the right call.
What to do instead
Testing your own project. In the Scratch editor use File, then Save to your computer to download the .sb3, then open it in Bilup with File, then Load from your computer. Many people work primarily in Bilup and upload the finished .sb3 back to Scratch. Keep backups when you do this.
Collaborating. Share the project on Scratch. It is fine to share unfinished work. If you want live editing with others, Bilup also has built-in collaboration that does not depend on Scratch at all.
Embedding. Share the project on Scratch, or download the .sb3 and use the Bilup Packager to turn it into a standalone file you can embed.
For tool developers
This section is for people building their own Scratch-related tools. To download a shared project you first fetch its metadata to get a project token, then use that token to fetch the data:
GET https://api.scratch.mit.edu/projects/<id>and read theproject_tokenfield.GET https://projects.scratch.mit.edu/<id>?token=<token>for the project JSON or SB3.
Browsers cannot call api.scratch.mit.edu directly because it does not send CORS headers, so browser-side code needs a CORS proxy; server-side code (Node.js) is not subject to CORS and can call it directly. For a ready-made downloader that handles all of this, see sb-downloader.