Skip to content

Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.

Studio: raise the startup bundle budget by #13998's growth (#14004)

Daniel Handbd6b3c78014
diff --git a/studio/frontend/scripts/check-bundle-budget.ts b/studio/frontend/scripts/check-bundle-budget.ts
index d9541588016..895b6c7b32a 100644
--- a/studio/frontend/scripts/check-bundle-budget.ts
+++ b/studio/frontend/scripts/check-bundle-budget.ts
@@ -191,8 +191,13 @@ export const BUDGET = {
// 377 B transfer (6,796,799 -> 6,798,138 B raw, 2,044,202 -> 2,044,579 B transfer against 1574001bc1, one
// machine and build, Node 24, chunk count unchanged at 95). Raised by that growth so main's margin stays where
// it was.
- transferBytes: 2_044_768,
- rawBytes: 6_798_298,
+ // Three Studio merges since add 3,117 B raw / 1,058 B transfer (6,798,138 -> 6,801,255 B raw, 2,044,534 ->
+ // 2,045,592 B transfer from 4ebc263832~1 to d4e005e73e, one machine and build, Node 24, chunk count unchanged at
+ // 95). Nearly all of it is searching a model maker in the Model hub (#13998, +3,038 B raw / +1,000 B transfer):
+ // its owner-to-uploads search lives in use-hub-model-search, a hook the startup screen already loads, so there
+ // is no separate module to lazy-load. #13999 and #13993 add 79 B raw / 58 B transfer. Raised by that growth.
+ transferBytes: 2_045_781,
+ rawBytes: 6_801_415,
};
// The chunk count is reported but not budgeted. Splitting a page out of the entry raises it while lowering the