Connect your own AI: Claude or ChatGPT working with your companies

ōrbita launches a connector for your own AI assistant. You paste an address into Claude, ChatGPT or another compatible assistant, authorize, and your AI reads the strategy, catalog, proposals and metrics of your companies and leaves drafts for you to review. It acts with your same permissions and never publishes or schedules (since October 10, 2026 it does, if you ask it to and confirm in the chat: see the entry from that date). On all plans.

Updated Author: Mónica Lorenzo

Areas: Your connected AIContentCalendar

What changes

If you already pay for an AI assistant, it can now work inside ōrbita. You connect it once and ask for things in its own chat: to rewrite the text of a post, to prepare drafts for a campaign, to explain the month’s metrics or to update a section of the communication plan. What it writes reaches ōrbita as a draft, in the same place where you work.

You bring the AI, with the subscription you already have. ōrbita brings what your assistant lacks: the real context of each company (strategy, brand voice, catalog, reviews, calendar and metrics) and the place where the work stays saved.

How you connect it

Update from October 10: with your permission and your confirmation in the chat, your AI now schedules posts and replies to comments and reviews, and sets up a new company with your documents: Your AI now schedules, replies and sets up your company.

Update from September 28: in Claude you no longer need to paste the address. ōrbita is in its connector directory and connects with one click; the steps are in Connect your AI.

You paste this address into your assistant, sign in to your ōrbita account and choose what you allow it:

https://mcp.orbitasolutions.org/mcp

You also have it in My account > Agents connected to your account, with a copy button. The guide Connect your AI explains the steps in Claude, ChatGPT, Claude Code and Cursor. In ChatGPT you connect it from developer mode, and which plans can write is decided by OpenAI.

With your permissions, and without publishing

Your AI acts with your identity: in each company it can do the same as you, never more. When you connect you see a list of permissions with checkboxes and untick what you do not want. And there are things it cannot do in any case: publish, schedule, tick on your behalf the declarations each network requires, see your billing or manage members.

Requests from ōrbita

On every post and every day of the calendar there is a new button, Ask my agent: you write what you want and the request waits. Your AI picks it up when you ask it to or, if it has scheduled tasks, when it is its turn to check them. ōrbita does not notify it: it is the assistants that ask.

You control access

In My account > Agents connected to your account you see every connected AI, when it was last used and which permissions it has, and you can remove a permission or disconnect it. In a shared company, in Connections you also see which people have an AI working in it and you can block someone’s without touching their access to ōrbita.

All releases

Published . Updated .