Repository navigation
Studio: pick the image and video model where the chat runs - #1159
Draft
sosidudku1 wants to merge 6 commits into
Draft
sosidudku1 wants to merge 6 commits into
sosidudku1 wants to merge 6 commits into
Conversation
…oaded Generate loads a downloaded local model by itself, so telling the user to load it first was wrong. Images and Video now say the model loads when Generate is pressed, or that it is not downloaded yet.
Studio named the machine "this Mac" everywhere: the local model group in the Images and Video pickers, the model row, the footer, the progress card, the mode hints and the first-frame notice. A Windows user saw "On this Mac · free, private" (ATO-1400). The words now follow the platform the bridge reports, the same rule as renderer.js THIS_MACHINE: "this Mac" on macOS and in the demo, "this computer" on Windows and Linux. "Show in Finder" follows the chat's fileManagerName() (Explorer / Folder). The mode-hint rule no longer matches on the hint's own words.
The Images and Video pages read the model catalog once per window. After the first local generation (which loads the model by itself) the footer still said "loads when you press Generate" and the picker "Not loaded" until the window was reloaded (ATO-1394). The shell now hands every non-core BR.studio.onLocal event to the pages' optional onLocal. A page reads the catalog again, quietly and once per burst, when a model finished loading (or failed to), a download finished, a model was removed, or one of its local jobs ended. Only the models' facts change: the chosen model, the form and the prompt stay.
This was referenced Nov 10, 2026
The Images and Video pages kept their own model choice, so with the chat on OpenRouter Studio still opened on a local FLUX it had been set to once (ATO-902). Owner decision 25.10: Studio always works where the chat does. - STUDIO.chatPlace() answers 'local', 'cloud' or null (settings not read yet) from selBackend() and rmNow(); a custom server counts as local, Fusion follows its orchestrator, as Improve prompt does (ATO-1304). - Images and Video open on a model of that place: the one last chosen there, else in the cloud the default, locally a loaded model, else a downloaded one; with nothing there, no model rather than the other place. Each place remembers its own choice (ATO-1305, ATO-1306). - When the chat has moved since the page was last shown, the page reads the catalog again and switches model, keeping the prompt. Never while a job is being made; it catches up when the job ends. - The Video first-frame hint asks to switch the chat to the cloud instead of picking a cloud model.
sosidudku1
force-pushed
the
feat/ato-1304-studio-follows-chat
branch
from
November 10, 2026 23:54
0615cde to
12f1bf5
Compare
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Studio kept its own model choice, unrelated to the chat. With the chat on Cloud · OpenRouter, Studio › Images still opened on a local FLUX it had been set to once, and asked to "load the model first".
Owner decision (25.10): Studio always works where the chat does. This PR:
STUDIO.chatPlace()instudio/shell.js:'local','cloud', ornullwhile the settings are not read yet (with an empty configselBackend()sayscloud, so that case is checked first). It readsselBackend()andrmNow()from renderer.js and does not change renderer.js. The pure rule (placeOf) is exported for tests. A custom server counts as local; Fusion follows its orchestrator, as Improve prompt does (main/studio-improve-prompt.ts). These two are proposals in the Linear issue and still need the owner's ok.Before the settings are read, both pages behave as before.
Why
Linear ATO-1304, ATO-1305, ATO-1306 (parent ATO-902).
Stack
Stacks on #1114 and #1157 and #1158 (same files). This PR's own change is the top two commits (the second answers review: Video keeps the latest model used in each place, a restored one included, the same way Images does).
How it was verified
studio-shell.test.mjs(cloud, local, custom, Fusion with a cloud or local orchestrator, empty settings),studio-images.test.mjs(default and remembered model per place),studio-video.test.mjs(pickModelper place, and the page switching with the chat and back, but not during a running clip; the latest model used in a place, a restored one included, is the one it comes back to). Not run yet.Not yet checked on the second Mac.