Skip to main content
To connect Jira, use an Atlassian connection (the same one Confluence uses, or a new one), then name the projects to read. It takes about five minutes.
Jira is in beta. It has been tested against recorded responses, not yet against a real account, so tell your MITHUNAI contact if a sync does not read what you expect.
Whatever the account can see, your assistant can quote to anyone who can ask it. Use an account that can browse only the projects you want read. Issue security levels inside Jira are not carried over. See Saved connections.

What you need

  • A Jira Cloud site, at an address like https://your-company.atlassian.net. Jira Data Center and sites on your own domain are not supported.
  • An Atlassian account that can browse the projects you want read.
  • The owner or admin role in your MITHUNAI organisation.

1. Create the API token in Atlassian

If you already saved an Atlassian connection for Confluence on the same site, skip to step 3: one connection reads both. Signed in to Atlassian as the account MITHUNAI will read as:
  1. Go to id.atlassian.com/manage-profile/security/api-tokens.
  2. Select Create API token. If you see a choice, choose the plain API token, not the one with scopes: MITHUNAI does not support scoped tokens yet.
  3. Name it so you will recognise it later, such as MITHUNAI.
  4. Choose an expiration date, from 1 to 365 days. Note the date: you will replace the token before then.
  5. Select Create, then Copy to clipboard.
Atlassian shows the token only once. If you lose it, create another.

2. Save it in MITHUNAI

  1. In the console, open Knowledge → Sources and choose Connect source.
  2. Under Issues and support, choose Jira.
  3. If you have no Atlassian connection yet, the form to add one is open. Fill in:
    • Connection name: such as Engineering Atlassian.
    • Site address: such as https://your-company.atlassian.net. It cannot be changed later.
    • Account email: the email of the Atlassian account the token belongs to.
    • Token: the API token you copied.
  4. Choose Save connection. The token is stored encrypted and is never shown again.

3. Choose what to read

  • Name: what the source is called in your list.
  • Projects: the short code at the start of each issue key, separated by commas. ENG-42 is in project ENG. At least one is needed.
  • Under Change filters: Resolved issues only reads only issues whose status is in Jira’s Done category, and Maximum issues is from 1 to 5,000, 500 by default.
Choose Connect source. The first sync starts straight away.

What is read

Each issue becomes one document: its key and summary, type, status, resolution and labels, its description, and its comments, oldest first. Answers cite it with a link to the issue. Left out: comments restricted to a role or a group, internal comments on service-desk issues, attachments, work logs, history and linked issues. No person is named: reporters, assignees, comment authors and @mentions are not kept, and email addresses typed into an issue are removed. An issue that has not been updated since the last sync is not read again.

When something goes wrong

Last modified on November 7, 2026