Skip to main content
Traditional Git hosting platforms struggle with large files. Most require Git Large File Storage (LFS) — a separate protocol that replaces large files with pointers and stores the actual content in a side channel. This adds complexity for developers and can break agent tooling. Mesa takes a different approach. Our storage layer is built from the ground up to handle large files natively. You commit them directly — no pointers, no LFS client, no special configuration. Large files are first-class objects in your repository just like any other file.

​
Key limits

There is no per-file size limit — the constraint is on the total pack size per push. See Limits Reference for the full list of system limits.

​
Recommendations

  • Commit large files directly. No LFS setup needed. Push them with git push like any other file.
  • Use MesaFS for agent workflows. The virtual filesystem streams file content on demand without downloading the entire repository.
  • Contact us if limits are too low. We can raise repository size limits for your org and are actively increasing these limits across the board. Reach out on Discord or email us.

​
What’s coming

We’re working on shallow clones (git clone --depth) and partial clones (git clone --filter) to make working with large repositories even faster. If either of these is important to your workflow, let us know.