Jira

Markdrop works with Jira Cloud in four ways:

• Create issues from what your team says and writes. A meeting's action item, decision or open question, a Team blocker or target, or any text in a page becomes a Jira issue. Markdrop AI drafts it; you check every field, then it's created as you.
• Tickets on Team. Each person's open Jira issues, what's overdue and blocked, and the team's issues by project, next to what they said in meetings.
• To-dos and issues stay in step. Tick a to-do that has an issue and the issue moves to Done in Jira. Move the issue to Done in Jira and the to-do ticks in Markdrop. Reopening works both ways.
• Ask and links. Ask "what's Ben working on in Jira?" or "status of PAY-123", and a Jira link pasted in a page shows the issue's live status.

Everything Markdrop does in Jira, it does as the person doing it, through their own Atlassian account. Jira's own permissions apply, and you are the reporter of the issues you create.

A workspace admin or the owner connects the workspace to one Jira site. Open Workspace settings and choose the Integrations tab. The Jira card shows one of three states:

To connect:

1. On the Integrations tab, select Connect Jira site.
2. Atlassian asks you to allow Markdrop to use your Jira. Pick the site and select Accept. You come back to the Integrations tab.
3. Choose:
   • Jira site, when your Atlassian account reaches more than one.
   • Default project for new issues and Default issue type. People can pick another project when they create an issue.
   • Projects on Team: which projects' issues show on Team. Pick as many as you like.
4. Select Save. The card says Connected.

Your own Atlassian account is connected as part of this. Everyone else connects theirs once (see Connect your Atlassian account).

Until an admin picks at least one project, Team shows nothing from Jira. The card says Pick projects for Team, and Team tells admins where to choose them. To change them later, select Change projects, tick or untick projects, and Save.

If the card says The admin connection stopped working, the Atlassian account that connected the site lost access (it was removed from Jira, or the access was revoked). Status updates from Jira pause until an admin selects Reconnect. Disconnect Jira… removes the connection. Issues stay in Jira and links stay on pages, but Team, status sync and chips stop until the site is connected again.

Markdrop creates and moves issues as you, so each person connects their own Atlassian account once:

• Settings › Account › Jira › Connect my Atlassian account, or
• Workspace settings › Integrations › Your Atlassian account, or
• the first time you create an issue: the dialog asks you to connect first.

Atlassian asks you to allow access. You come back to where you started, connected. Disconnect in Settings › Account removes your connection. Issues you made stay in Jira.

Markdrop finds who each workspace member is in Jira: by their own connection, otherwise by their Markdrop email address, when Jira shows emails. That's how a meeting's "Ben will…" becomes an issue assigned to Ben, and how Team knows whose tickets are whose.

You can create an issue from:

• A meeting's decisions, action items and open questions. Point at the item and select + Jira next to it (on a phone it's always there). You need to be able to edit the meeting page.
• Any text in a page. In Edit mode, select text and choose Create Jira issue (the ticket button) in the toolbar, or open a block's ⋮⋮ menu and choose Create Jira issue.
• A line on Team. A blocker, target, action item or "working on" line: open its ⋯ menu and choose Create Jira issue….

Markdrop AI drafts the issue from the item: a short Summary, a Description, the Issue type, a Priority when the item sounds urgent, the Assignee when the item names its owner, a Due date when it gives one ("by Friday"), and a few Labels. Check and change every field, pick another Project if you need to, then select Create issue.

The issue's description gets your text, the item quoted, and a link back: "From Checkout launch sync · Nov 6, 2026 at 12:34", which opens the meeting at that moment of the transcript (or the page at that block). The item then shows a chip: the issue's key, status and assignee.

When AI can't run (it's off for the workspace, or today's AI budget is used up), the draft is the item as written, and the dialog says why. Creating the issue never depends on AI.

On a meeting page, select Jira issues in the meeting bar. Markdrop drafts an issue for every open action item that doesn't have one yet. Untick the ones you don't want, edit the summaries, choose the project, and select Create 2 issues (the number you kept).

For a meeting restricted to its attendees, the dialog warns you: the link back only opens for people who can open the meeting in Markdrop, but the text you put in the issue is visible to everyone who can see the issue in Jira. Edit the summary and description before you create it.

A to-do or item with an issue shows a chip after it: the key (PAY-4), the status (To do, In progress, Done) and the assignee. Select it to open the issue in Jira.

What you see depends on your own Jira access. If your Atlassian account can open the issue, you see its status and assignee. If it can't, or you haven't connected, you see the key only. Markdrop never shows you an issue you couldn't open in Jira.

A Jira link pasted in a page (https://your-site.atlassian.net/browse/PAY-123) gets the same live chip, for people whose Jira can open it.

For a to-do with an issue:

From Jira, the to-do is changed as the Markdrop member who changed the issue, or else as the person who linked it, when they can edit the page. If someone is editing the page at that moment, it's changed when they're done.

If Markdrop can't move the issue (you haven't connected your Atlassian account, or Jira doesn't let you make that change), nothing moves, and the chip shows a warning. Point at it to see why, and select Retry once it's fixed.

With Team on and at least one project on Team, the board shows each person's Jira issues: assigned to them, in those projects, updated in the last 30 days.

• Each card: how many are in progress, to do and overdue, and their top three open issues.
• Jira under Team at a glance: the team's counts, the overdue and blocked issues, and open, overdue and recently done issues per project. Picking a group narrows it to the group.
• A person's page: their issues grouped as Overdue, Blocked, In progress, To do and Done lately.

An issue is overdue when its due date has passed and it isn't done, and blocked when its status name says "Blocked", it has a "blocked" label, or it's flagged.

Details on Team come through your own Atlassian account, so you only ever see issues your Jira lets you open. Until you connect, Team shows counts only (how many to do, in progress, overdue, blocked and done lately per person, in the chosen projects), never a key, summary or link, with Connect to see the tickets.

Team refreshes from Jira every few minutes when you open it; status changes in Jira show up sooner.

Ask Markdrop reads Jira when your question is about tickets: "what's Ben working on in Jira?", "which tickets are blocked?", "status of PAY-123". It uses your own Atlassian connection, so it only sees what you can open in Jira, and it cites the issues with links. If you haven't connected, Ask answers from Markdrop only.

With the hosted MCP server, an AI app can check the connection (jira_status), draft and create issues from an item on a page or meeting (create_jira_issue: draft: true returns the AI draft and creates nothing), link an existing issue to an item (link_jira_issue), list the team's issues by person, project or status (list_jira_issues), and move an issue to done, to do or in progress (transition_jira_issue). The same rules apply: issue details only through your own Atlassian connection. Connecting is done on the website. See MCP tools.

• Markdrop stores each person's Atlassian tokens encrypted, and uses them only for what that person does or sees.
• Only what you can read is sent to AI or to Jira: the item, its section and the meeting's or page's title and date.
• A chip, Team details and Ask answers show an issue only to people whose own Jira can open it.
• Markdrop keeps a short-lived copy of the issues it shows (key, summary, status, assignee, due date) to keep pages fast.
• Disconnecting your account deletes your tokens. Disconnecting the site stops Team, chips and status sync for the workspace.

Drafts use Claude Haiku. Each draft is one short request; "Jira issues" for a whole meeting is one request. Markdrop caps what Jira drafts can cost per day and per month across the server; when the cap is reached, drafts are the item as written until the next day or month. AI follows your workspace's AI setting: with AI off for the workspace, drafts are the item as written.

This is for whoever runs the Markdrop server, once. Workspaces can't connect Jira until it's done.

1. Go to the Atlassian developer console and select Create › OAuth 2.0 integration. Name it "Markdrop" and agree to the terms.
2. Permissions › Jira API › Add, then Configure. Under classic scopes, add read:jira-work, write:jira-work and read:jira-user. Add manage:jira-webhook too. Markdrop asks for offline_access itself.
3. Authorization › OAuth 2.0 (3LO) › Add. Set the Callback URL to https://markdrop.in/integrations/jira/callback (your server's address, followed by /integrations/jira/callback).
4. Distribution: set Distribution status to Sharing, fill in the vendor and privacy details, and save, so people outside your own Atlassian organization can connect.
5. Settings: copy the Client ID and Secret into the server's environment as MARKDROP_ATLASSIAN_CLIENT_ID and MARKDROP_ATLASSIAN_CLIENT_SECRET, and restart it. MARKDROP_TOKEN_ENCRYPTION_KEY must be set too.

Jira's status updates reach Markdrop through webhooks that Markdrop registers on the site when an admin connects it, and renews before they expire. Nothing else is needed in Jira.