Tabs
Parallel AI conversations and generators, one tab each
Tabs lets you run several independent AI conversations side by side, each wired to a preset generator: general chat, image, video, document writer, text-to-speech, or a custom endpoint. Tabs show live status and unread badges, keep their own chat history and Output panel, and a Run All button sends one prompt to every tab. The whole tab set is saved inside the current project and follows you across devices.
Who it is for: Anyone who wants to fan a prompt out to several assistants or generators and compare results in one place.
Quick answers
Is a tab an embedded website or chat site?
No. Each tab is a self-contained conversation with one generator preset; Tabs posts your prompt to a server endpoint and shows the result. Nothing is embedded in frames.
Can I pick a different AI backend per tab?
Not with the floating backend selector in the current build; it is ignored here. A Custom tab can be pointed at any endpoint that speaks the same streaming chat format, for example a local service exposed through TektBridge.
Where are my tabs stored?
In a hidden .tabs-state.json file inside the selected project's workspace, including full chat histories and links to generated files. Switching project switches the tab set.
Why does Run All break my image or TTS tab?
Run All sends a plain chat request to every tab regardless of its preset. Trigger generator tabs individually.
Does Save to Workspace work?
Not in the current build. The button targets a workspace route that does not exist and reports success anyway. Use Download and add the file through the Hub.
Loading the user manual…