Lobu is now in ChatGPT

ChatGPT becomes much more useful at work when it can understand the company around the question you are asking. The customer, the project, the decisions that have already been made, the work that is still open and where all of that information came from.

Until now, connecting Lobu to ChatGPT meant enabling developer mode, creating a custom connector, pasting an MCP URL and scanning for tools. The new ChatGPT listing makes that connection much more direct.

You can now connect Lobu directly to ChatGPT through the published Lobu app. Open the listing, follow the connection prompts and sign in to your Lobu workspace. No custom connector or MCP URL required.

The more interesting part, though, is not the easier setup. It is what ChatGPT can work with once Lobu is connected.

Bring your company context into the conversation

A customer question rarely lives in one system. The promise might be in an email. The discussion is in Slack. The remaining work is in Jira. The release itself is in GitHub.

You can find each of those pieces yourself and paste them into ChatGPT, but then you are still doing the work of connecting everything before the AI can help. And the next time somebody asks a related question, much of that context has to be assembled again.

Lobu connects those sources into a shared model of the company: customers, projects, requirements, decisions, tasks and the relationships between them. The context stays in your Lobu workspace rather than belonging to one AI conversation, and authorized agents can work from the same company knowledge with links back to the original sources.

That distinction matters. Giving ChatGPT access to another source is useful. Giving it company context that can persist across conversations and agents is a different thing.

Take a customer launch. The team has agreed on a date in Slack and the release is merged in GitHub. At first glance, everything looks ready. But the contract also requires SSO acceptance and that test is still open in Jira.

Looking at the release alone could produce the wrong answer. The useful answer comes from understanding that the open Jira task is connected to a requirement for the same customer and therefore still affects whether the launch is actually ready.

That is the kind of context Lobu is designed to preserve. Once the relevant sources are connected, ChatGPT can retrieve the customer, their requirements, the latest conversations and the outstanding work as part of the same question.

Ask from where you’re already working

Not every task needs to start automatically. Sometimes you already know what you need and simply want to ask. With Lobu connected, you can do that directly from ChatGPT without gathering the company context yourself first.

For example:

Use Lobu to prepare a brief for my next customer call. Find the latest conversations, commitments and unresolved issues. Include the source links.

Or:

Use Lobu to check whether this project is ready to launch. Look for open requirements and blockers, identify their owners and tell me what information is missing.

The difference is that you’re not starting from a blank conversation and manually rebuilding the context around the task. ChatGPT can use the company context available through Lobu.

And when the work is ongoing, you don’t need to keep prompting it. Lobu teammates can work around the responsibilities you’ve given them, using their instructions, permissions and tools to move work forward without waiting for you to initiate every step.

If you are starting fresh, create a Lobu workspace and connect the sources you want your agents to work with first.

Your company memory should not belong to one AI client

A useful conversation often produces something worth keeping. Maybe a decision was made, an assumption was corrected or somebody clarified what needs to happen next.

Normally, that knowledge risks staying inside the conversation where it happened. With Lobu, you can ask ChatGPT to save the useful result back to the shared workspace. Another authorized agent can then retrieve that context later, whether you are working in ChatGPT, Claude, Codex or another Lobu client.

That means changing the AI or interface you are using does not have to mean starting again. The interface can change while the company context stays with the company.

Let specialist agents handle the workflow

ChatGPT also does not have to find every piece of information and perform every individual step itself. You can ask it to use a specialist agent your team has already set up in Lobu.

That agent can have its own instructions, permissions and tools for a particular workflow. Instead of explaining the process again each time, Lobu can hand the task to an agent that already knows what it is responsible for and can complete that workflow end to end within the boundaries you have set.

The same controls still apply when you reach that agent through ChatGPT. Lobu permissions and approval requirements are not bypassed just because the request started somewhere else. You can inspect the shared context in Lobu and follow its sources back to the original tools.

Some workflows can also use a connected browser or computer through the Lobu Chrome extension or Mac app. Those connections still need to be set up and authorized separately before an agent can use them.

Connect Lobu in ChatGPT

Open the Lobu listing in ChatGPT, complete the connection prompts and start a conversation with Lobu selected. The ChatGPT setup guide has the steps and a few prompts to try.

If you already have a Lobu workspace, try asking a question you would normally have to open several tools to answer. If you are new to Lobu, create a workspace and connect the sources you want it to work with first.

The published ChatGPT app connects to Lobu Cloud. Lobu remains open source, and self-hosted instances can still connect through their own MCP endpoint.

The important part is that your company context does not have to be rebuilt inside every AI tool you use. Lobu keeps that context shared and persistent, and ChatGPT can now work from it directly.