Storyboard AI

Make a video with the Rovo agent

The Storyboard AI agent for Rovo builds and renders a narrated video from a Jira work item through chat, without you opening the editor. It's available on any Jira site that has Rovo enabled.

Starting a conversation

Find Storyboard AI in Rovo chat, or mention the agent from a work item. It opens with three suggested prompts:

  • Draft a storyboard from this issue

  • List the attachments on this issue

  • What's the render status?

You can also just describe what you want — for example, "make a 5-scene walkthrough from the screenshots on this ticket, use Sarah's voice, LinkedIn Post template."

The agent reads the work item's description and attachments before it builds anything, so how the work item is written shapes what you get back. See Preparing a Work Item for the Agent for the conventions it recognizes — scene-to-attachment mapping in particular.

What the agent can do

Area

What you can ask for

Read

Summarize the current storyboard; list the work item's attachments; list available voices

Build

Draft scenes from the work item's content, a topic prompt, or per-scene input you supply

Edit

Rewrite narration, add or delete a scene, reorder scenes, attach an image or clip to a scene

Style

Change template, aspect ratio, voice, or caption style

Produce

Generate narration audio, start a render, report render status

Remove

Delete a single scene or the whole storyboard

"Frame" and "scene" mean the same thing — one segment of the video.

Where the video shows up

The agent writes into the same storyboard the editor uses, so anything it builds appears in the Storyboard AI panel on that work item. When a render finishes, the MP4 is available from that panel.

Open the panel to confirm a finished render. The agent reports that it started the render, but the finished video is recorded when someone opens the work item's Storyboard AI panel. If you ask the agent for status and it still says "rendering" well after you'd expect, open the panel on the work item — that settles it.

Permissions

The agent acts as you, not as itself. Every action is checked against your own Jira permissions on that work item: you need Browse Projects to read a storyboard and Edit Issues to change or render one. If you cannot edit the work item, the agent cannot edit its storyboard on your behalf.

The first time the agent acts for you, Rovo asks you to grant consent. Until you do, actions fail with a permission error.

Confirming an action

Some actions ask you to confirm before they run. If a confirmation prompt appears empty or without buttons in the chat thread, open the Agent sessions row for that conversation and confirm there instead.

Limits

  • The agent works on one Jira work item at a time — the one in context.

  • It never adds files to the work item's Attachments section — not the finished MP4, not anything else. The video is delivered through the Storyboard AI panel. Files move one way only: the agent can pull an existing attachment into a scene, never push one out.

  • It can only use files that are on the work item itself. A file you uploaded in the editor is copied to Attachments only if it's under roughly 4 MB, so a larger clip — most videos — won't be visible to the agent. Drag it onto the work item's Attachments section if you need the agent to use it.

  • It cannot resize video. If you ask it to attach a clip larger than 1920×1080, it refuses and tells you to add that clip from the editor, which resizes automatically. See FAQs.

Last updated: