You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
NEXT_PUBLIC_SELF_HOST env var: when true, exposes the /create upload flow for users running their own GPU backend. Default (unset/false) shows the "Coming soon" placeholder used for splattie.app.
gsplat (Apache 2.0) replaces diff-gaussian-rasterization (INRIA, non-commercial) as the rasterizer in the LAM fork. Output is byte-identical for our use case.
cpu and gpu uv extras in backend/pyproject.toml. uv sync --extra cpu for the FastAPI server, uv sync --extra gpu for the full LAM pipeline.