Please expose an official HTTP (or MCP) API for Grok Bot desktop so third-party / companion UIs can:
List bots / agents for the signed-in account
Create a bot (name, title, description)
Send a message to a bot and read replies / transcript (poll or stream)
Without needing Windows UI Automation against the Grok Bot window.
Why this matters
Today, companion apps (e.g. a clearer local shell over Grok Bot) can only drive the desktop UI via accessibility APIs. That works, but:
Chromium/Grok Bot briefly steals window focus on create/send (Invoke), which feels unprofessional even when we restore focus afterward
UI Automation breaks when Grok Bot’s UI labels/layout change
There is no stable contract for create/chat—only workarounds (UIA) plus partial surfaces like webhook wake and remote MCP tools, which do not replace create-bot or chat I/O
An official API would let us keep Grok Bot as the runtime while shipping a better UI layer with zero focus stealing and fewer breakages.
Suggested shape (illustrative)
Authenticated as the same Cursor account as the desktop app:
GET /v1/bots
POST /v1/bots{ name, title?, description? }
POST /v1/bots/{id}/messages{ text }
GET /v1/bots/{id}/transcript?since=…
Or equivalent MCP tools with the same semantics.
Out of scope / fine to defer
Full theming of the stock Grok Bot chrome; local stdio MCP; anything that bypasses account auth.
Platform
Windows Grok Bot desktop (also useful for macOS/Linux companions later).
Happy to share a concrete companion use case and reproduce the focus-steal issue if helpful.
This is a slightly different request but on parallel with the obvious void in Grok’s services you are pointing out.
There is no way for Grok Bot -a multiplatform interface with transcription and live chat- to drive Grok TUI. You can give it ‘always allow’ access on a particular laptop, but if you have an existing coding session ongoing on that device, you can’t have Grok Bot simply ‘take over’, to use the better Grok Bot interface - say starting on the laptop, then moving to your phone and interfacing from there while you’re on the move, all the while having Grok Bot drive Grok TUI on the laptop while you are physically away from the device.
I just assumed when Grok Bot was released, that it didn’t need a ‘Connector’ like it does for e.g. Gmail, to manage it’s own product from it’s own company. It goes without saying that of course they should integrate with each other, but they don’t.
Hey, thanks for the detailed feature request. We don’t often get a request with ready-made endpoint shapes and clear motivation, and it really helps.
There’s currently no official public API or SDK for Grok Bot for this use case (list/create bots, send message, read transcript). So yes, UI automation is basically the only option right now. The focus-stealing and breaking on layout changes is a real downside of that approach.
I’ve passed this along to the team as a feature request. I can’t share a timeline or make promises about the exact interface shape, but the request is logged in the right form.
If you can share a concrete use case and steps to reproduce the focus-steal issue, that’d be really helpful. Please post it in this same thread. Also, if you can, tell us which of the four operations (list / create / send / transcript) is most critical for you. That’ll help us prioritize.