How do I get my AI worker to pull all the files reliably instead of hitting upload or fetch failures?

Give your AI worker the crawl token instead of a pile of uploads. The token pulls every chapter, module spec and file straight from the Free Builder server, so there is nothing to upload and nothing to fetch by hand.

1. Get the AI crawl token and prompt from free-builder.com. It comes with the free kit.
2. Hand the token to your worker at the start of the session, before it writes any code.
3. Let it crawl the whole kit first: 12 modules and the reference codebase.
4. Turn off file uploads as the source. The token is the source.
5. If a fetch fails, have the worker retry the token fetch, not a pasted file. The files live on the server.
6. Keep the scaffold in place. It ships frontend, backend, database, auth and payments so the AI builds inside known boundaries.
7. Check the 12-module spec at /kit-documentation/ when something looks off.

The kit costs nothing. You keep every customer and every dollar. The paid part is optional: the $19 AI employees material. No rush, but every evening you spend wrestling uploads is an evening you did not spend selling.

Mouldi