Orkeste agents read your Jira, Linear and Trello boards, spot the work that has stalled and write the status update your team actually reads, then post it to Slack or Confluence. You decide which of those actions wait for a person to approve them.
The real state of a project is spread across Jira tickets, Linear issues, a spec in Confluence and a thread in Slack. The weekly meeting turns into people reading their own tabs out loud.
A ticket sits untouched for eight days, a dependency slips, an owner is on leave. Nobody notices until standup, when the deadline is already at risk.
Someone senior spends Friday afternoon copying ticket counts into a summary and rewriting it for three audiences. That is judgement time spent on assembly.
Each agent has a short briefing and only the connectors it needs. From the Starter plan, you can chain agents in a swarm so one hands its output to the next.
Runs every Monday at the hour you set. Reads the open issues in Jira or Linear, groups what moved, what stalled and what is due, and posts a short digest to the project channel in Slack.
Looks for issues with no recent activity or past their due date, and adds a comment asking the owner for an update, so the question lands where they already work.
Reads requests that arrive by email and turns each one into a Trello card or a Jira issue with a clear title, the context and a suggested priority.
Writes the decisions taken in a meeting to a Confluence or Notion page from the notes you paste, so the record exists before anyone forgets.
The first agent above, as it runs every week.
Priorities, trade-offs and the conversation with a stakeholder whose project is late. The agents gather and draft; the people who own the plan decide what changes in it.
No. They take over the collection and the first draft: reading boards, spotting stalled work, writing the digest. Your project managers keep the decisions, and any action you mark as sensitive waits for their approval.
Jira, Linear, Trello, Confluence, Notion, Airtable and Slack, among the 26 connectors. Each connector page lists exactly what an agent can read and change. In Jira, for example, an agent creates issues and adds comments but does not change an issue status or assignee.
Yes. Each agent can run every hour, every day at a set hour, or every week on a set day and hour, in addition to the runs you start yourself.
Each run uses credits from your plan, from 2 credits with Claude Haiku 5.5 to 235 with Claude Fable 5.1 depending on the model you choose, and you can set a daily and a monthly budget per agent.