Skip to content
Retain7

Retain7 for ChatGPT

Give ChatGPT access to your team's project memory

ChatGPT is where many teams think through architecture, and none of it reaches the tools that touch the code. Decisions get re-explained to every agent, and agent findings never make it back to the chat.

How it works

One memory, shared across your tools

Add a custom connector in developer mode with the server URL and OAuth, then approve the sign-in. ChatGPT can then look up what your coding agents saved and file new decisions into the same project memory.

Saved in ChatGPT

The public API is versioned in the URL.

Found in Cursor

The next session searches before it answers, so the decision is already in its context.

An example memory, not customer data.

Setup

Add Retain7 to ChatGPT

A minute or two, once. The same steps are in the docs.

  1. In ChatGPT, add a custom connector (developer mode) with this URL and OAuth authentication, then approve the sign-in.

    https://retain7.com/mcp
  2. So every session checks the team memory before it answers, add this line to your ChatGPT custom instructions (Settings, Personalization):

    This team keeps its project memory in Retain7. Before answering how this project does something, call the retain7 memory_search tool.

Questions about ChatGPT

What do I need to enable in ChatGPT?

Custom connectors in developer mode. Add the server URL on this page with OAuth authentication, then approve the sign-in.

Does ChatGPT train on my memories?

Retain7 does not use your memories to train models, and its search vectors come from a model on its own servers. What ChatGPT does with data it reads follows OpenAI's terms.

Can ChatGPT write to the memory?

Yes, if your role allows saving. New memories can wait in the review inbox until an owner or admin approves them.

Can read-only teammates use ChatGPT here?

Yes. Viewers can search and read, but nothing they connect can save.

Give ChatGPT a memory that lasts.

Free for 14 days on the Team plan. No card needed.