- Table of contents
1. What is the Planio MCP Endpoint?
In short, MCP acts as a secure bridge between your Planio account and an AI assistant of your chosing.
The Planio MCP Endpoint lets your AI assistant connect to your Planio account using the Model Context Protocol (MCP).
It implements a remote HTTP MCP endpoint for Redmine, providing access to the Redmine API via the MCP in a way that AI assistants can understand and use.
Once connected, an AI assistant of your choice can read or update project data based on the permissions you grant. This setup makes it possible for your assistant to work directly with your project information, for example to read and summarize issues, create new tasks, or update project information.
2. Your Feedback is Important
We’re rolling out this new feature as a Beta so you can explore its potential, experiment with different assistants, and decide what works best for your team. The setup is fully optional and remains under your control at all times.
Because this is an early release, your feedback is important and useful to us. We’d love to hear about your experience, what works well, and where we can improve as we continue refining this feature. This guide will walk you through how the integration works, how to get started, and what to keep in mind when using an AI assistant with your Planio data.
3. Getting Started: What do you need to prepare?
Your AI assistant will be accessing your Planio data using the Planio API, so you need to have this set up already, as covered here. For full API reference, please see the Planio API docs and Redmine documentation.
Making the decision to allow an AI assistant to access and/or edit your data in Planio is a decision that only an administrator can make. Therefore to be able to enable the MCP for your account an administrator must login and set this up:
• Navigate to Administration → Settings → API
• Put a tick in the box "Enable the MCP Endpoint"
• Decide whether your AI assistant should be able to just read your data or also make changes and tick the box "Read-only mode" accordingly
• Click Save. If read-only mode is enabled, the AI assistant can view and summarize data but cannot make changes in your account like create, update, or close issues or edit wiki pages.
There are two ways to connect with the MCP endpoint:
• API token: Use your personal API key if your AI assistant allows setting a custom HTTP header (for example, Claude Code).
• OAuth 2.1: Connect through OAuth with dynamic client registration.
If you’re unsure which option to use, OAuth is recommended for most users because it allows fine-grained permissions and can be revoked at any time.
When using OAuth, you can decide which permissions the AI assistant should have. You can adjust these at any time by editing the granted scopes under My Account → Authorized Applications in Planio. With an API token, all actions are performed exactly as the connected user, and there’s no way to restrict them via scopes. With OAuth, access is limited to the scopes you choose.
Once you have these ready, you can move on to setting up the connection between your AI assistant and Planio.
In general, you will need:
• Remote MCP endpoint URL: https://your-account.planio.com/mcp (swap out your-account.planio.com with your actual Planio domain name)
• Your personal API token (found in the sidebar of the "My Account" section in Planio)
• Your personal Planio log in details
Please note, each AI assistant has a different way to set up the Planio MCP connector and so it's best to ask the AI assistant how to do that. You should get a step by step guide and answers to any difficulties you may encounter.
★ Remember ★ Not all AI assisants allow connecting to an MCP endpoint for free. You may have to sign up for a paid plan to do so.
4. Revoking Access
You can revoke an AI assistant’s access at any time by:
• Disabling the MCP endpoint in Administration → API
• Revoking the OAuth application under My Account → Authorized Applications
5. Example Set Up in Claude
Here we will go through connecting a commercially available AI assistant, in this case Claude, with a Planio account as an example:
• Go to your AI assistant → Settings → Connectors → Add custom connector
• Use https://your-account.planio.com/mcp as the Remote MCP endpoint URL, replacing your-account.planio.com with your actual Planio domain
• Click Connect
• You will be redirected to log in to your Planio account
• Choose which actions your AI assistant should be permitted to perform and click Allow
• You will be redirected back to your AI assistant where you can just get started asking your AI assistant questions
6. What can I use the MCP Integration for?
Here we will give some examples of what you can do with your AI assistant in Planio.
★ Remember ★ Always review changes and answers suggested or made by an AI assistant, especially when write access is enabled.
Common Prompts
• Show me the available Planio projects
• Show me recent activity in project XY
• Create a new issue in project XY with title A
• Update the status of issue #123 to "Resolved"
Clear Project Status Breakdown
With many issues at different phases it is useful to get an overview of the current status of your project in a clear, easy-to-read format. An AI assistant can do this in seconds for you:
Summarizing Issue Content
Some project issues have been worked on for months or longer, and it's a lot of work to go through these and get a view of what has happened, what's even relevant and what's not. You can let your AI assistant go through all the relevant points in the issue and give you a short, concise summary keeping you up to date with less effort.
Reducing Repetitive Work
Let's say you work for a recruiting company and receive multiple CVs per email. Instead of going through each CV and using different programs to open each one, let your AI assistant extract the information and consolidate it into easy to read Planio issues.
And voilà, your issues have been created in Planio, assigned and ready to move to the next step of the recruitment process.
Why not take it one step further and have it compare the applicants, giving suggestions as to who could be the best fit and why:
Creating Issues from Meeting Notes
No one wants to admit it, but taking meeting notes by hand is still happening and typing that up and making actionable items out of it afterwards is an arduous task that can be made easier with an AI assistant.
Just upload your hand written meeting notes to your AI assistant and ask it to create a wiki page of the notes.
The AI assistant can also create issues for the actionable items that arose in the meeting.
Editing a Wiki Page
You can ask the AI assistant to update a wiki page:
7. Hints, Tips and Troubleshooting
→ Can I use an AI assistant with Planio if I am not an administrator?
Yes, so long as an administrator has set up the functionality in your account, non-administrators can also connect an AI assistant to their Planio account. The AI assistant will then only have access to the projects and data that you have access to.
→ What happens with my data if I enable the MCP endpoint? What does the AI company see?
Enabling the MCP endpoint does not share or upload any data to AI companies. Planio itself never sends your data anywhere. Data is only shared once you explicitly connect or authorize the AI service to fetch it, strictly limited to the permissions of the currently authenticated user. Once authorized, the assistant has access to the data through the endpoint and may access it when responding to your requests.
Planio continues to store all customer data exclusively on its own servers.
→ I have an Enterprise account, why can I not see these new settings?
If you want to use an AI assistant with Planio and you have an Enterprise account, we can set that up for you personally. Please contact us directly.
→ How can I know if the AI assistant made the change or if a user did?
All API actions are attributed to the user account whose API token is being used. So from Planio's perspective the user makes the changes, not the AI assistant. There will not be a note added to specify that the AI assistant made the change on a user’s behalf.
• If you see it in the UI: It would show that "Anne User" updated the issue
• In reality: The action could have been triggered by the AI assistant using Anne User's API credentials via the MCP endpoint
It is possible of course, to create a dedicated “AI user” account and use that account’s API credentials for the AI assistant.
→ Why is my assistant saying it can't perform a simple task?
The MCP endpoint can perform the same actions that the API can. The API can only do what the user account itself is allowed to do. There are no extra permissions for the AI assistant.
→ Why is the answer taking so long, why is it so slow?
The response time from an AI assistant can vary a lot! Some common reasons it takes a long time is that the type of request made means looking through a lot of issues and making hundreds of API calls. This isn't the best use of the MCP endpoint so try to avoid such requests. Examples would be "Find all the duplicate issues in my account" or "Which issues have the biggest attachments?". Sometimes it is slow because it gets stuck without a certain permission, which you may have to manually allow. You can often click "Always allow" which will speed it up in the future.
→ Why did the AI say something changed when nothing changed in Planio?
If the answer implies that a change was made in Planio and this is not the case, please contact us so we can take a closer look.
→ Does the AI assistant see all projects or only those I can access?
The AI assistant can only access the same projects, issues, and data that the connected user account has permission to see.
→ How can I get more accurate responses?
Follow this general cheat sheet for good AI requests:
• Use action verbs,
• Be specific,
• Be explicit,
• Give one task at a time.
For example:
• Ineffective : "Can you review all open bugs, prioritize them, update statuses, and notify the team?"
• Effective : Split it into 3 direct requests; 1. “List all open bugs in project X”, 2. “Suggest priorities for these issues”, 3. “Update issue #123 to priority High.”
8. Current Limitations
As this is a Beta, there are still some limitations that we'll gradually address over time.
To prevent surprises, here's what you can't yet ask your AI assistant to do in Planio:
• Fill in custom fields
• Assign groups
• Add or edit time entries
• Upload attachments
• Analyze, edit, or read attachments that aren't available in full text
Please let us know if you encounter any other limitations, so we can improve the integration, add support for them, or document them here for others.
No credit card required. Sign up only takes 60 seconds.