A user asks OpenAI to let ChatGPT chats and Dots message each other directly

The request in plain terms

A user on community.openai.com posted a feature request asking for native, two-way communication between an existing ChatGPT conversation and a Dot. The poster wants messages to flow in both directions, under permissions the user sets, and ideally arrive in near-real time.

The post does not define what a Dot is, and we will not guess beyond what it shows. In this account, a Dot is a separate assistant that can be given access to a shared document and can act within capabilities the user has authorized. ChatGPT and the Dot do different but complementary jobs on a long-running private project.

The poster's complaint is simple: the two assistants cannot coordinate without the user copying information from one to the other.

A workaround built on Google Docs

The poster has not been waiting idle. They built a working stopgap: both ChatGPT and the Dot have access to one shared Google Doc, which serves as a message board. Each can read what the other wrote, add new entries, post task updates and keep shared background about the project.

The poster says it works surprisingly well. That matters, because it shows the need is practical and not hypothetical. The poster offers to share more technical detail about the setup privately if OpenAI wants it.

It is also a clever, low-tech pattern. A shared document needs no special integration, and the user can open it to see exactly what the assistants have told each other.

Where the document falls short

The post lists the limits of the approach:

• No real-time contact. Each side has to open the document, or check it on a schedule, to see anything new.
• No notifications. Writing a message does not wake up or alert the intended reader.
• No delivery receipts. Nothing says whether a message was received, read or acted on.
• Broken continuity. When a ChatGPT conversation reaches its session limit, the link to its successor has to be rebuilt.
• Leftover overhead. The user still spends time managing communication that could happen directly.

In short, a shared document is a bulletin board, not a messaging system. Someone has to keep walking over to look at it.

What a native version would need

The poster sketches the feature in some detail. A ChatGPT conversation could send a message straight to a designated Dot, and a Dot could send messages or notifications back. Messages would arrive immediately when possible, or wait in a queue until the recipient is next active.

The request also asks for basic message status and acknowledgments, and for communication that survives across successor conversations within the same Project. Even a simple shared inbox with delivery and acknowledgments, the poster says, would be a big step up from checking a connected document.

The poster also draws a line on safety. The user would decide which Dots and conversations may talk to each other and with what permissions. A receiving assistant could act on a message using abilities it already has, while consequential actions and new permissions would still need the user's approval.

A wish, not a feature

Nothing in the post suggests OpenAI is building this. The poster ends with a direct question: is there already a supported way to connect ChatGPT conversations and Dots, perhaps through a connector, an API or another mechanism? If not, the poster asks that it be considered for the roadmap. The post shows no reply from OpenAI.

So the status is open. We can't say whether a supported route exists, and we can't say whether the company has this on any plan. Treat the post as a well-described example of a common pain point and not as news of a coming feature.

The permission-first design is the most reusable part. If you ever connect several assistants, the same questions apply: who can message whom, what can they do on their own, and what needs a human yes?