To connect Claude to project management in AlianHub, open AI, then Connect your AI, copy the address shown, and add it as a custom connector in Claude. ChatGPT works the same way through a connector, and Claude Code connects with a personal token. Once connected, you ask in everyday words and your AI finds, adds and changes work as you, with every change logged.
This guide covers how it works, the setup steps, what your AI can and cannot do, and the safety controls behind it.
How connecting your AI works
AlianHub ships an MCP server. MCP is the standard that lets AI apps such as Claude, ChatGPT and Claude Code use tools in other software. When you connect, your AI app talks to that server and acts inside your workspace.
Four rules shape everything:
- It works as you. It sees what you can open and changes only what your role lets you change.
- Only while its app is open, or while it runs on a schedule you set up in that app.
- Every change is logged in the audit log, under its name and yours.
- Undo, or approve first. Some changes happen at once with Undo. Wider ones wait for your approval.
So if you cannot open a project, neither can your AI. It never widens what you can see.
Before you start
You need an AlianHub account first. There is no page where people make their own account. You get one in one of two ways:
- An owner or admin of your workspace invites you, and you choose a password from the invitation link.
- You install AlianHub yourself, and the install wizard creates the first workspace with you as its owner.
Your AlianHub address must also be reachable from outside your network, because your AI app reaches AlianHub over the internet.
Tip: Connecting by address is on for a new server. On a server set up earlier, an owner can switch it on under Settings, Instance, AI. If it is off, use a token until then.
Step-by-step setup
On the Connect your AI page
- In AlianHub, open AI in the left menu, then Connect your AI. When you first join a workspace or finish the install wizard, this page appears as "One last step". You can skip it and come back later.
- Pick your app and follow its steps. Copy the address with the Copy button.
- The status line at the top changes to "Connected" when your AI makes its first call, and shows when it was last seen.
- A first prompt appears: "Set up my project". It asks a few questions and shows a plan. Nothing is made until you say yes.
The page also lists what your server allows and names anything your admin has switched off on that install.
In Claude
Open Settings, then Connectors, and choose Add custom connector. Paste the address. Claude opens AlianHub and asks what it may do.
In ChatGPT
Open Settings, then Apps & Connectors. Turn on Developer mode under Advanced settings, then create a connector and paste the address. ChatGPT offers this on some plans only.
In Claude Code or another tool
Choose Create a token on the Connect your AI page. Under My account, choose New token, name it, pick the projects it may use and tick what it may do. The token is shown once, so copy it then. Your admin may require an expiry date.
The consent screen and permissions
When your AI app opens AlianHub, you see a consent screen asking whether to allow the app to act for you, with a list of what it asks to do. You choose the one workspace it may use, then Approve or Deny. If your workspace has not approved the app yet, you can ask an admin and wait for their decision.
Some permissions are granted together when you approve. Others start unticked, so you tick only what you want.
| Permission | What it allows | Default |
|---|---|---|
| tasks:read | Read tasks you can see | Granted on approve |
| tasks:write | Create and change tasks for you | Granted on approve |
| projects:read | Read projects you can see | Granted on approve |
| docs:read | Read pages and documents you can see | Granted on approve |
| time:read / time:write | Read time logs and log time for you | Granted on approve |
| tasks:manage | Manage tasks, such as setting the person and date | Starts unticked |
What your AI can do
Out of the box, your AI can:
- Find and read your tasks, projects, lists, docs, comments and time entries.
- Add tasks and subtasks, comment, attach links and log time.
- Change a task's status to In progress or In review.
- Show you where something is, with a link.
With the extra permissions, and unless your admin has switched the matching tools off, it can also change, assign, move and archive tasks several at a time, write docs, read chat in the channels you are in, and work with tags, task links, lists, doc comments and goals. Requests for new fields, saved views, project setups, new projects and automation rules always wait for approval.
Things to say
A few examples of everyday requests and what happens:
- "Move Design to In Progress." Applied at once, with Undo.
- "Add a task Call supplier to the Launch list." Applied at once, with Undo.
- "Show me all my tasks across every project." Nothing changes; you get a link.
- "Set bulk 01 to bulk 20 to high priority and assign them all to Priya." Waits for approval as one request.
- "Show me this list grouped by Stage." Waits for approval, because it saves a view everyone on the project sees.
Open the screen you mean in AlianHub first, because "here" and "this list" refer to the screen you last opened.
Your AI app can also offer ready-made asks by name: Set up my project, Plan my day, What is at risk, Write the status report and Triage what is new. Each shows you a result first and changes something only after you say yes.
What it cannot do
These limits are the core of the safety model:
- It cannot delete a task, a doc, a comment or a project.
- It cannot remove people, change who may do what, or touch billing.
- It cannot read your direct messages.
- It cannot close a task on its own authority. A close follows your project's rule.
- It cannot change or switch off an automation rule that already exists.
- It does nothing while its app is closed.
Delete, merge and deploy are absent from the tool list, not merely disabled. Admins also have one switch to pause every connected agent, and you can disconnect your app whenever you like. For autonomy levels and audit views, see AI controls. Admins setting up a server should read the admin guide.
FAQ
Does connecting Claude give it access to everything in AlianHub?
No. It acts as you, so it sees only the projects you can open and changes only what your role allows.
Can ChatGPT connect to AlianHub?
Yes. Turn on Developer mode in ChatGPT, create a connector and paste your address. ChatGPT offers custom connectors on some plans only.
What happens if my AI makes a mistake?
Single-task changes come with Undo, wider changes wait for your approval, and everything is recorded in the audit log.
Can I use Claude Code with AlianHub?
Yes. Create a personal token on the Connect your AI page, scope it to chosen projects and tick only the permissions it needs.
Try it yourself
Read the full Connect your AI guide, explore the AI features, or open the live demo and try a few everyday requests before you connect your own workspace.