- Home
- Skills
- Agents & Orchestration
- Livingboard
More screenshots
Works with the AI Agents you already use
Livingboard
Sketch on a shared whiteboard. Your agent reads it, draws back editable diagrams and tables, and keeps board and code in sync.
Free
Livingboard
Example session with this skill installed
Open a whiteboard called agensi-demo, draw how Livingboard works, and add a table of its packages.
- Read your context and instructions
- Compiled the livingboard
Summary
Opened board agensi-demo at http://localhost:3333/?p=personal_projects-668520&board=agensi-demo and added two editable drawings: a diagram of how Livingboard works and a table of its packages. The first screenshot on this page shows the result.
What the agent did
| Step | Tool | Result |
|---|---|---|
| 1 | open_board | Created the board in .livingboard/boards/agensi-demo.ydoc, readable copy in agensi-demo.md |
| 2 | board_overview | Empty board, effort medium, no drawings yet |
| 3 | draw_diagram | drawingId: d_brjiy7t02p2g, 18 elements |
| 4 | draw_table | drawingId: d_g6j8n8hs2p81, 19 elements |
| 5 | get_board_image | Checked the render, 1020×587 px |
On the board
How Livingboard works
flowchart LR
B[Browser board] <-->|Yjs live sync| S[Livingboard server]
S <-->|MCP| A[Coding agent]
S -->|saves| F[(.livingboard/boards)]
A -->|edits| C[Project code]
Packages
| Package | Job |
|---|---|
| shared | Document schema, types, Yjs keys |
| server | Sync, segmentation, MCP, persistence |
| web | Excalidraw, AI overlay, panel, views |
| livingboard | npm package: bundled server and CLI |
In the repo
Livingboard wrote this readable copy next to the board, so it shows up in diffs:
# agensi-demo
> Livingboard board of personal_projects, written by Livingboard on every change: edit the board, not this file (`/whiteboard agensi-demo` in Claude Code).
## Claude's drawings
- **How Livingboard works** (diagram): Browser board · Livingboard server · Coding agent · .livingboard/boards · Project code · Yjs live sync · MCP · saves · edits
- **Packages** (table): Package · Job · shared · Document schema, types, Yjs keys · server · Sync, segmentation, MCP, persistence · web · Excalidraw, AI overlay, panel, views · livingboard · npm package: bundled server and CLI · Packages
Next steps
- Change a cell in the table: the agent receives it as
drawingEditsand updates the project. - Write
@claude ...next to a drawing: the answer appears on the board. - When the project changes, the agent redraws the same drawing with
replace.
Connects securely to your tools. The creator never sees your data.
About this skill
Livingboard is an open-source whiteboard your coding agent can read and draw on. You sketch screens, flows and ideas; the agent reads them, builds, and puts back on the board what it did.
This skill tells the agent when and how to use the board. The board itself is the free Livingboard MCP server (MIT), which you install separately. The skill shows you the commands and asks before installing anything.
What it does
- Reads what you draw: sketches, handwriting and arrows, through the agent's own vision.
- Draws back: screens, tables, charts and Mermaid diagrams, made of normal Excalidraw elements you can move and edit.
- Keeps the board current: every drawing has an ID, so the agent redraws it in place instead of piling up copies.
- Treats your edits as requests: change a cell in the agent's table and the agent updates the project to match.
- Lives in your repo: boards are saved in
.livingboard/boards/, with a readable.mdcopy that shows up in diffs.
How it works
- The agent checks that the
open_boardtool is available. If not, it shows you the install commands and what they change, and waits for your yes. - It opens the board and gives you the URL.
- It follows the working rules the server returns, and draws what it builds or decides.
- It reads your changes with
wait_for_changesand updates the project.
When to use it
Best for screens, flows, architecture and plans, where position carries meaning. For bugs, refactoring and precise rules, text is still faster: the board gives the chat a map, it doesn't replace it.
About the screenshots
The first screenshot is the board from the demo session on this page. The others come from the project's README: a real board, with the agent's side scripted for the recording through the same MCP tools.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires installing the free livingboard npm package. Tested on Claude Code only. One agent per board at a time. Sketch recognition depends on the agent's vision. Many large images make a board heavy to load. Arrows drawn by the agent are not bound to shapes. The board page loads a font from the esm.sh CDN.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 12 days ago
- Free with an Agensi account
- Software
- livingboard MCP server
Node.js 22+A web browser
Permissions
File Scopes
Shell: only for the install commands, after the user says yes. Read/write files: the agent updates the project to match the board; the server saves boards in .livingboard/ and its data in ~/.livingboard/; on Claude Code, `livingboard install` edits ~/.claude/settings.json and adds a skill. Browser: the board opens in a tab. Network: npm install, a local server on 127.0.0.1, and the Excalidraw font from esm.sh. Board content stays on the computer unless --lan is turned on.