Cubelife
Give a long-running coding agent a live pixel-art avatar that updates with what it is doing so you can glance at status without reading logs.
Overview
CubeLife is an MCP server for the Build phase that syncs a persistent pixel-art agent character to your agent’s live state in real time.
What is this MCP server?
- Persistent pixel-art character tied to agent activity in real time
- stdio MCP server (`cubelife` npm package, v0.3.2)
- Dashboard-issued `CUBELIFE_API_KEY` and `CUBELIFE_AGENT_ID` env vars
- Optional companion site at life.cubeworld.co.za for viewing the character
- MCP package version 0.3.2 via npm identifier `cubelife`
- stdio transport with 2 required environment variables
- Source published under Ivy-Innovation/cubelife (cli subfolder)
What problem does it solve?
Long agent runs feel opaque—you cannot tell at a glance whether the bot is thinking, blocked, or finished without reading verbose transcripts.
Who is it for?
Indie builders who demo agents, stream coding sessions, or want a fun persistent mascot while Claude Code or Cursor works in the background.
Skip if: Teams that need formal incident timelines, PagerDuty-style alerts, or headless CI-only agents with no human watching.
What do I get? / Deliverables
After wiring the stdio MCP server with your dashboard credentials, your agent can update a visible character so session status is obvious without digging through logs.
- stdio MCP integration pointing at npm `cubelife`
- Live-updating pixel character linked to your registered agent
- Repeatable env-based auth for local agent sessions
Recommended MCP Servers
Journey fit
Agent-facing UX and presence layers sit in the build phase when you wire tooling around Claude Code, Cursor, or Codex workflows. CubeLife is registered as an MCP stdio server (npm `cubelife`) that agents call to sync character state—not app frontend code.
How it compares
Agent presence visualization via MCP—not structured logging, not a generic chat skill.
Common Questions / FAQ
Who is CubeLife for?
Solo and indie builders running MCP-capable agents who want a real-time visual indicator of what the agent is doing.
When should I use CubeLife?
Use it during build-phase agent workflows when you are actively monitoring or showcasing an agent and want state reflected on a persistent pixel character.
How do I add CubeLife to my agent?
Create an agent in the CubeLife dashboard, set `CUBELIFE_API_KEY` and `CUBELIFE_AGENT_ID`, and register the npm stdio server `cubelife` (v0.3.2) in your MCP client such as Claude Code.