# Playcal

> Playcal is an open-source MCP server for games and play. Reliable edit-settle-verify loop for AI agents working in the Unity Editor. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, self-hosted
- **Website:** https://pypi.org/project/playcal/
- **Repository:** https://github.com/rizgarozan/playcal
- **MCP Registry name:** io.github.RizgarOzan/playcal
- **Categories:** games and play, developer tools
- **Listing updated:** 2026-11-14

## About

Playcal targets the boring failures that make agents unreliable in a game editor: acting before compilation or domain reload has finished, Play Mode sitting on frame 1 because the editor is unfocused, a modal dialog blocking everything, or reporting a prefab as done when it was never saved. Its loop waits for the editor to settle, checks the change actually landed, shows only what changed and then asserts.

Around that loop sit scene diffs that survive an Editor restart, performance and screenshot gates, flaky-test quarantine, logs, replay, and bot playtests calibrated against real players, so a result like a bot finding level 12 easy can be compared with human data. It is Unity first, built on Unity's official CLI, with Godot working headless and Unreal postponed.

playcal mcp turns each module into a tool such as playcal_ingest, playcal_logs and playcal_loop. The project calls itself pre-alpha, and needs Python 3.10 or later.

## Features

- **Edit-settle-verify loop:** Waits for the editor to settle and checks that a change landed.
- **Compact observation:** Returns only what changed to save tokens.
- **Scene diffs:** Diffs that survive an Editor restart.
- **Performance and visual gates:** Baselines for perf and memory plus screenshot checks.
- **Calibrated bot playtests:** Compares bot results against real player data.
- **Flaky-test quarantine:** Isolates unreliable tests.

## Good for

- Let an agent edit a Unity scene and confirm the edit saved
- Gate a change on performance or screenshot regressions
- Compare bot playtest difficulty with real players
- Keep editor logs small enough for an agent to read

## Install

### Claude Code (Terminal)

```sh
claude mcp add playcal -- uvx playcal
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "playcal": {
      "command": "uvx",
      "args": [
        "playcal"
      ]
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "playcal": {
      "command": "uvx",
      "args": [
        "playcal"
      ]
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "playcal": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "playcal"
      ]
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "playcal": {
      "command": "uvx",
      "args": [
        "playcal"
      ]
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "playcal": {
      "command": "uvx",
      "args": [
        "playcal"
      ]
    }
  }
}
```

## Tools

- `playcal_ingest`: Ingests playtest events.
- `playcal_logs`: Reads editor logs.
- `playcal_loop`: Runs the edit-settle-verify loop for editor writes.

## FAQ

### Which engines are supported?

Unity first; Godot works headless; Unreal is postponed.

### What does it need?

Python 3.10 or later, installed with pipx install playcal.

### How stable is it?

The project describes itself as pre-alpha.

### How do I add the MCP server?

claude mcp add --transport stdio playcal -- playcal mcp.

### Is Playcal free?

Yes. Playcal is open source, so you can use and self-host it for free.

### How do I add Playcal to Claude Code?

Run this command in your terminal: claude mcp add playcal -- uvx playcal

### What are the best alternatives to Playcal?

The closest alternatives to Playcal are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Concord MCP](https://ai-powered.com/mcp/concord-mcp): Concord MCP is a freemium MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/playcal
