We're excited to introduce StoriesOnBoard MCP — our Model Context Protocol server that allows AI agents such as Claude Code, Cursor, and Claude Desktop to work with StoriesOnBoard through a standard MCP interface.

MCP (Model Context Protocol) is an open standard that lets AI agents connect to external tools and systems in a structured, reliable way. Instead of working only from the text in a chat, an MCP-enabled agent can discover data, read context, and perform actions through clearly defined tools.

In practice, this means AI agents can do more than just answer questions. They can work with real product data, understand the current state of a system, and help complete tasks inside the tools teams already use.

For StoriesOnBoard, MCP makes it possible for agents to explore story maps, read card details, discover valid reference data, and now also create and update cards. This opens the door to more practical workflows where AI can support backlog grooming, story map maintenance, and day-to-day product work in a much more direct way.

With StoriesOnBoard MCP, agents can explore story maps and read card data using the existing tools:

In addition to the read access, MCP tools that allow AI agents to perform real write operations in StoriesOnBoard:

Together, these tools build on the existing MCP read capabilities and create a natural agent workflow:

ListStoryMaps → discover story maps

GetStoryMapData → discover collaborators, colors, statuses, annotations, priorities, and personas needed for valid write inputs

GetStoryMapCards → inspect the activity / task / subtask hierarchy and releases

CreateStoryMapCard → create a new card

UpdateStoryMapCard → update an existing card

GetStoryMapCardDetails → verify the result

This makes StoriesOnBoard MCP suitable not only for reading project context, but also for letting AI agents actively help manage and maintain story maps.

In practice, StoriesOnBoard MCP can help AI agents support a range of product workflows across reading, creating, refining, and syncing work.

For example, teams can use it to turn a PRD into a first draft of a story map, identify missing stories in an existing flow, rewrite stories into a consistent format, suggest effort scores for unestimated work, generate release summaries from completed items, or sync a selected slice of work into Jira.

These are the kinds of workflows where MCP becomes especially useful: not just reading story map data, but helping teams actively shape, improve, and move work forward.

We extended StoriesOnBoard MCP well beyond reading and editing card content. Agents can now plan releases, reorganize the board, and build personas — covering the parts of product work that used to require the UI.

Your AI agent can now run the full release lifecycle on a story map without you touching the UI, using three new tools:

Created and edited releases behave exactly like UI-made ones, with board activity, real-time updates on both boards, team-chat notifications, and integration sync.

With the new MoveStoryMapCard tool, your agent can rearrange the structure of a story map, not just read and write card content. A single call can reorder activities, reorder or reparent tasks across activities, and reorder, reparent, or reschedule subtasks — including scheduling a subtask into a release or pulling it back to unscheduled.

Because spatial position in story mapping carries meaning — priority, sequence, grouping — this lets an agent reshape the map the way you actually think about it. Moves are validated all-or-nothing, detect no-ops, and fire the same real-time updates, notifications, and DevOps sync as a manual drag.

Your agent can now build and maintain personas directly from research material instead of leaving you to type them in:

The net effect: an agent can derive personas from a transcript, attach them to the right stories, and keep them aligned over time without any manual UI work.

StoriesOnBoard MCP has graduated from a selected-customer beta to general availability. MCP is no longer gated behind manual per-workspace access — it's now a standard part of the Standard, Pro, Plus, Premium, and trial plans, so any team can connect their AI agents.

To reach GA we finalized and locked down the public tool inventory, with safety annotations on every tool, and ListStoryMaps now returns aggregate counts (activities, tasks, subtasks, releases) per map so agents can quickly size a workspace.

When you copy content out of the rich-text editor and paste it into a plain-text destination — Slack, Jira plain fields, issue trackers, Notepad, notes apps — the result now keeps the structure you saw in the editor by serializing as markdown.

Lists, tables, headings, blockquotes, code blocks, mentions, card links, and embeds all carry over in their markdown form, and inline formatting (bold, italic, strikethrough, inline code, links) survives too. Copying part of an ordered list now stays a properly renumbered ordered list instead of collapsing into separate paragraphs. And because the copied markdown is the editor's own on-save format, it pastes back cleanly into other StoriesOnBoard markdown fields. Pasting into rich destinations like Google Docs, Word, or rich Slack is unchanged — so you get clean formatting everywhere.