August 31
Manage story map annotations with your AI agent
Annotations are the little markers — an icon and a color, or an emoji — you use to classify and visually tag cards on a story map. Until now your AI agent could assign existing annotations to cards, but creating a new one meant leaving the agent flow and doing it by hand in StoriesOnBoard. That gap is closed: seven new MCP tools let your agent manage the whole annotation vocabulary of a map.
Create, update, archive, restore — and clean up
Your agent can now:
List every annotation on a map — including archived ones — with its name, look, position, and how many cards currently use it.
Create new annotations with a name, an icon (chosen from a catalog of 112 icons) and one of eight colors.
Update any of those — rename, restyle or reorder — with the response telling you exactly what changed.
Archive an annotation to retire it from new use while every existing card keeps its marker, and restore it later with all assignments intact.
Permanently delete an annotation when it should really go away. This is deliberately guarded: the agent must pass an explicit confirmation flag, and the deletion removes the marker from every card and the map in one atomic step, reporting how many cards were affected.
It fits the flow you already use
The end-to-end agent workflow now runs without a single context switch: create a story map, define the annotations the source material calls for, create the cards, and tag them — all from your chat. Everything shows up live on open boards, and each operation is recorded on the Board Activity Panel, clearly marked as an agent action, so your team always sees what the AI did.
Card links in your agent's results
When your agent creates, updates or moves a card, the result now includes the card's direct StoriesOnBoard link. Ask your agent for a summary of what it just built and you get working links you can click or paste into Slack — no more hunting for the card by name. Links respect your permissions as always: opening one requires access to the board.
More reliable MCP connections
Two under-the-hood improvements for everyone connecting AI agents:
Your agent stays signed in. MCP sign-ins are now backed by stable, centrally managed credentials, so our deployments and infrastructure changes no longer risk logging your agents out early. A connected agent stays connected for the full lifetime of its session.
Stricter MCP clients now connect cleanly. Some MCP clients that validate the OAuth handshake very strictly could fail to connect due to a subtle mismatch in our server's discovery metadata. That's fixed — if your client refused to connect before, give it another try.