Linear in ChatGPT: The App and What It Can Do

Yes, Linear works in ChatGPT, in two ways. The simple one is the Linear plugin in ChatGPT’s plugin directory: open Plugins, find Linear, install it and sign in to your Linear workspace. OpenAI describes it as finding and referencing issues and projects, for status updates, themes in recent work and progress on a feature. The fuller one is to add Linear’s own MCP server, https://mcp.linear.app/mcp, as a custom app in ChatGPT’s developer mode. That gives ChatGPT the tools Linear publishes for finding, creating and updating issues, projects and comments, and ChatGPT asks you before each change. Either way it works with your Linear permissions, and what you can connect depends on your ChatGPT plan and, in a company workspace, on your admin.

Names here have moved more than once. What ChatGPT called connectors became apps, and apps now sit inside plugins; Linear’s own page still calls its listing a connector. The steps below follow OpenAI’s and Linear’s documentation as it reads at the end of September 2026. For the full list of what Linear’s server can reach, see what the Linear MCP server does; this piece is about ChatGPT.

Route 1: the Linear plugin

OpenAI’s plugins documentation↗ (opens in a new tab) describes one directory shared by ChatGPT and Codex, across the web, desktop and mobile. Installing works the same way for every plugin:

1. In ChatGPT, open the Plugins tab and search for Linear.
2. Select the plus button to install it.
3. Follow the prompt to connect Linear: sign in with the Linear account you want ChatGPT to use and approve access to the workspace.
4. Start a new chat, then name Linear in your message or pick it with @, so ChatGPT uses it rather than guessing.

OpenAI’s listing frames the plugin around reading: using your Linear issues and comments to write status updates, spot themes and explain where a feature stands, with attachments not included. One of its example prompts asks ChatGPT to triage or update issues, so some changes may be offered; whether they are depends on the actions the app exposes, your admin’s settings and your Linear permissions. The listing adds that availability depends on the plugin, your plan and workspace settings, and that some connections need an admin to set them up.

Linear’s ChatGPT integration page↗ (opens in a new tab) describes the older shape of the same thing: a connector you add from ChatGPT’s settings and use as a source in deep research, keeping each person’s Linear permissions. It lists Plus and Pro, outside the EEA, Switzerland and the UK, and Team, Enterprise and Edu. It mentions no creating or editing of issues. If the two pages disagree, trust what ChatGPT shows your account.

Route 2: Linear’s MCP server in developer mode

Developer mode turns ChatGPT into a full MCP client for any remote server you add. OpenAI’s developer mode guide↗ (opens in a new tab) lists it for Plus, Pro, Business, Enterprise and Education accounts on the web, with support for all tools, read and write, and calls it powerful but dangerous: meant for people who understand prompt injection, mistaken writes and malicious servers.

Linear’s MCP documentation↗ (opens in a new tab) gives the server address and says it uses OAuth 2.1 with dynamic client registration, the sign-in ChatGPT handles on its own. Linear lists setup for Claude, Cursor, VS Code, Codex and others, but not for ChatGPT, so the steps below are OpenAI’s generic ones for any remote server:

1. In ChatGPT on the web, open Settings, then Security and login, and turn on Developer mode.
2. Go to Plugins and use the plus button to create a developer-mode app. Name it Linear and enter the server URL below.
3. Choose OAuth for authentication, create the app, and sign in to Linear when it opens. Approve the workspace.
4. Review the tools ChatGPT discovered. You can switch individual tools off in the app’s settings.
5. In a new chat, add the app from the + menu under Developer mode.

https://mcp.linear.app/mcp

Those are OpenAI’s steps for connecting a server↗ (opens in a new tab), which note that availability depends on your account and workspace policies. On a work account, a missing Developer mode switch is a question for your ChatGPT admin, and ChatGPT connectors not working covers the rest of the usual failures.

Read-only first

Linear offers a read-only address that only ever exposes read tools. Adding it as a second app, or instead of the first, is the safest way to try Linear in ChatGPT:

https://mcp.linear.app/mcp/readonly

Linear also accepts a Linear API key, including one restricted to read, in an Authorization: Bearer header. OpenAI’s developer mode guide lists OAuth, no authentication and a mix of the two, not a custom header, so in ChatGPT use OAuth or the read-only address instead.

What it can read and change

Through the MCP server, ChatGPT gets what Linear documents: tools to find, create and update issues, projects and comments, with Linear adding more over time. Linear does not publish a tool-by-tool list, so the reliable list is the one ChatGPT shows in the app’s details after you connect. Deleting is not documented, so do not plan around a delete tool you have not seen there. A good first session reads only:

• “Using Linear, list the tools you have, then tell me which Linear workspace and teams you can see.”
• “Summarise the issues in the Payments project that changed this week: what shipped, what is in progress, what is blocked.”
• “Read every comment on ENG-231 and list the open questions.”

Then writes, one at a time: “Draft issues in the Platform team for each action in these meeting notes. Show me the titles and descriptions first and create nothing until I say so.” Ask ChatGPT to give the identifier of every issue it created or changed, so you can check them in Linear.

Approvals: when ChatGPT asks first

• In developer mode, write actions need your confirmation by default. ChatGPT treats any tool not marked with the readOnlyHint annotation as a write, so how often it asks depends on how Linear’s server labels its tools. You can remember an approval for the rest of that conversation.
• For plugin apps, OpenAI documents App permissions that decide when ChatGPT asks. Depending on the app and workspace, the options include Always ask, Any changes, Important actions and Never ask. For Linear, Any changes is a sensible setting: reads go through, changes wait for you.
• In a Business, Enterprise or Edu workspace, OpenAI’s plugin controls↗ (opens in a new tab) let admins turn apps on by role and, where an app supports Action control, allow read-only actions or a custom set. OpenAI advises starting every connected service with read actions.

None of these settings adds access. They only change when ChatGPT stops to check. The outer limit is always your own Linear account.

Limits to know

• It acts as you. Linear treats the signed-in user as the actor, so issues and comments ChatGPT writes appear under your name. If your team needs to tell them apart, agree a habit, such as a line at the end of each comment.
• One workspace per sign-in. Linear says each workspace needs its own authentication, so two workspaces means two apps.
• Developer mode is web only. The plugin works on web, desktop and mobile.
• Private teams stay private. The server sees only what your Linear account can see.
• Uninstalling a plugin does not disconnect everything. OpenAI notes that separately connected MCP integrations stay connected until you disconnect them in ChatGPT.
• Prompt injection is real here. Issue descriptions and comments are written by anyone in your workspace and sometimes by customers, and ChatGPT reads them as text it may act on. Keep write approvals on.

A board ChatGPT joins with a role

Linear plans a lot: cycles, projects, initiatives and releases. If what you need is one list of work that people and ChatGPT share, fenbs is a smaller board with four lanes, To Do, Next Up, In Progress and Completed, and three kinds of task, feature, enhancement and bug. ChatGPT connects to https://fenbs.ai/api/mcp the developer-mode way above, holds your role narrowed by the scopes you tick, and every change is recorded under its name. Be aware that fenbs sets no tool annotations yet, so ChatGPT treats every fenbs tool as a write and asks you to confirm each call. fenbs has no cycles, sprints, due dates or Linear import. The steps are on connecting ChatGPT to fenbs, and the fair comparison is fenbs vs Linear.