[PR #12667] [CLOSED] perf: Prevent model list API call from blocking UI #9796

New Issue

GiteaMirror · 2025-11-11T18:31:42-06:00

GiteaMirror commented

2025-11-11 18:31:42 -06:00

📋 Pull Request Information

Original PR: https://github.com/open-webui/open-webui/pull/12667
Author: @thearyadev
Created: 4/9/2025
Status: ❌ Closed

Base: dev ← Head: async-model-list

📝 Commits (2)

9f228d2 fix: prevent model list api call from blocking ui
df274e3 fix: refactor models.subscribe use

📊 Changes

2 files changed (+28 additions, -6 deletions)

View changed files

📝 src/lib/components/chat/Chat.svelte (+22 -0)
📝 src/routes/(app)/+layout.svelte (+6 -6)

📄 Description

Changelog Entry

Description

The API call to /api/models can take a long time. the ui waits until the api call promise is resolved before continuing any other ui state change or conducting any other api call. user sees a blank screen until the request has completed. This PR allows for the models list api call to run without blocking the UI.

This delay can become large if there are many model api urls, or if the servers are offline/are responding slowly.

Changed

set models store asynchronously to prevent UI blocking
src/lib/components/chat/Chat.svelte onMount subscribes to models and runs the initNewChat function when models changes, ensuring the UI is hydrated with the new data.

Additional Information

tested by adding an artificial delay to the endpoint /api/models in backend/open_webui/main.py:

@app.get("/api/models")
async def get_models(request: Request, user=Depends(get_verified_user)):
    import asyncio 
    await asyncio.sleep(10)
    ...

before:

this video demonstrates the blank UI while the call to /api/models completes

https://github.com/user-attachments/assets/a26ad416-38c0-4606-91b9-5108185b1bd3

after:

when the api call is done without blocking, the user gets a quick load, and is able to interact with the ui while the request completes in the background.

https://github.com/user-attachments/assets/1b03d5df-4593-4ba1-81e8-b0e33f0e942f

user navigates during model load functions as intended.

the initNewChat function is not called if the user navigates. The model list is still refreshed once it becomes available.

https://github.com/user-attachments/assets/b509e839-bd7a-44a8-9b78-86dbd2885367

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/open-webui/open-webui/pull/12667 **Author:** [@thearyadev](https://github.com/thearyadev) **Created:** 4/9/2025 **Status:** ❌ Closed **Base:** `dev` ← **Head:** `async-model-list` --- ### 📝 Commits (2) - [`9f228d2`](https://github.com/open-webui/open-webui/commit/9f228d2c81a166c37894bf942fa02a1283fa1960) fix: prevent model list api call from blocking ui - [`df274e3`](https://github.com/open-webui/open-webui/commit/df274e35340b6a4b1c7aa61445c4564258cae98c) fix: refactor models.subscribe use ### 📊 Changes **2 files changed** (+28 additions, -6 deletions) <details> <summary>View changed files</summary> 📝 `src/lib/components/chat/Chat.svelte` (+22 -0) 📝 `src/routes/(app)/+layout.svelte` (+6 -6) </details> ### 📄 Description # Changelog Entry ### Description The API call to `/api/models` can take a long time. the ui waits until the api call promise is resolved before continuing any other ui state change or conducting any other api call. user sees a blank screen until the request has completed. This PR allows for the models list api call to run without blocking the UI. This delay can become large if there are many model api urls, or if the servers are offline/are responding slowly. ### Changed - set `models` store asynchronously to prevent UI blocking - `src/lib/components/chat/Chat.svelte` onMount subscribes to `models` and runs the `initNewChat` function when `models` changes, ensuring the UI is hydrated with the new data. --- ### Additional Information tested by adding an artificial delay to the endpoint `/api/models` in `backend/open_webui/main.py`: ```python @app.get("/api/models") async def get_models(request: Request, user=Depends(get_verified_user)): import asyncio await asyncio.sleep(10) ... ``` #### before: this video demonstrates the blank UI while the call to `/api/models` completes https://github.com/user-attachments/assets/a26ad416-38c0-4606-91b9-5108185b1bd3 #### after: when the api call is done without blocking, the user gets a quick load, and is able to interact with the ui while the request completes in the background. https://github.com/user-attachments/assets/1b03d5df-4593-4ba1-81e8-b0e33f0e942f #### user navigates during model load functions as intended. the initNewChat function is not called if the user navigates. The model list is still refreshed once it becomes available. https://github.com/user-attachments/assets/b509e839-bd7a-44a8-9b78-86dbd2885367 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

GiteaMirror added the pull-request label 2025-11-11 18:31:42 -06:00

GiteaMirror closed this issue

2025-11-11 18:31:42 -06:00

GiteaMirror referenced this issue

2026-04-19 21:48:49 -05:00

[GH-ISSUE #9796] Categories in leaderboard and server-side search #15651

GiteaMirror referenced this issue

2026-04-25 05:14:49 -05:00

[GH-ISSUE #9796] Categories in leaderboard and server-side search #31179

GiteaMirror referenced this issue

2026-05-05 16:06:53 -05:00

[GH-ISSUE #9796] Categories in leaderboard and server-side search #54316