WazziCode Godot
MCP server that connects AI coding assistants to a live Godot editor, enabling scene, node, script, resource, runtime, debugger, and test workflows through natural language.
README
WazziCode Godot
WazziCode Godot connects Codex and Claude—plus other MCP-compatible AI coding clients—to a live Godot editor. It exposes scene, node, script, resource, runtime, debugger, and test workflows through the Model Context Protocol (MCP).
This project is an MIT-licensed derivative of hi-godot/godot-ai. See NOTICE.md for the exact upstream base, retained compatibility names, and update guidance; the dated candidate analysis is in docs/UPSTREAM-EVALUATION.md.
Quick start
Prerequisites: Godot 4.5+, Node.js 20+, either uv or Python 3.11+, and Codex, Claude, or another MCP client.
git clone https://github.com/robertobendi/wazzicode-godot.git
cd wazzicode-godot
node ./bootstrap.mjs "/absolute/path/to/your-godot-project"
Then open the target project in Godot, enable WazziCode Godot under Project > Project Settings > Plugins, and use the WazziCode Godot dock to configure Codex, Claude, or both. The same MCP capabilities are available to each client.
For platform-specific paths, prerequisites, what the bootstrap changes, and manual recovery, follow SETUP.md.
What it can do
- Inspect and edit scenes, nodes, properties, resources, scripts, signals, UI, animation, materials, particles, cameras, and environments.
- Search and modify project files with structured diagnostics and undo-aware editor operations.
- Run projects, inspect editor/runtime state, capture screenshots, read logs, profile performance, and execute Godot tests.
- Route calls to a selected editor session when several Godot projects share one backend.
- Keep large tool catalogs manageable through per-domain rollups and read-only
godot://resources. - Configure Codex, Claude, and other supported MCP clients from the editor dock.
See docs/TOOLS.md for the maintained tool and resource catalog.
Architecture
Codex / Claude / another MCP client
|
local stdio attach bridge
|
Python FastMCP HTTP backend
|
authenticated WebSocket
|
Godot editor plugin (GDScript)
The control path is local by default. Anonymous telemetry is also off by default and requires explicit opt-in; see docs/TELEMETRY.md.
Product name and compatibility identifiers
The visible product name is WazziCode Godot. The following mature upstream identifiers intentionally remain unchanged because launchers, saved editor settings, self-update logic, and MCP clients already depend on them:
- Python import:
godot_ai - Python distribution and legacy CLI:
godot-ai - Godot add-on path:
addons/godot_ai - editor settings namespace:
godot_ai/* - health/protocol names such as
/godot-ai/statusandgodot://...
A source installation also exposes the friendly CLI alias wazzicode-godot; both entry points run the same compatibility-preserving implementation.
Development
Create the development environment:
.\script\setup-dev.ps1
./script/setup-dev
Run the Python checks from the repository root:
.\.venv\Scripts\python -m pytest
.\.venv\Scripts\ruff check .
./.venv/bin/python -m pytest
./.venv/bin/ruff check .
Godot integration tests require a Godot editor binary and the repository test project; use script/ci-godot-tests in a compatible shell. These are validation commands, not claims about the current checkout's test result.
Provenance and upstream updates
WazziCode Godot currently starts from Godot AI 3.1.1 at upstream commit b1d767657fac87ef2a13a02c8ed58f4aa441db48. Keep upstream as a separate remote so attribution and future merges stay reviewable:
git remote add upstream https://github.com/hi-godot/godot-ai.git # once
git fetch upstream
git log --oneline HEAD..upstream/main
Merge or rebase upstream only on a dedicated branch, then re-apply and verify WazziCode branding, opt-in telemetry defaults, and bootstrap behavior before publishing. More detail is in NOTICE.md.
Status
This repository is a source derivative of the mature Godot AI lifecycle. Do not assume a WazziCode-specific PyPI, Godot Asset Library, or marketplace release exists unless this repository publishes the corresponding release artifact.
The inherited in-editor updater still verifies and installs official hi-godot/godot-ai releases. In a WazziCode source checkout, update with Git and rerun the bootstrap; using the upstream updater intentionally switches the add-on back to upstream Godot AI. The inherited publishing workflows are repository-guarded so this fork cannot accidentally publish the upstream godot-ai package.
License
MIT. The original copyright and license are retained in LICENSE and in the distributable add-on. Attribution and modification notes are in NOTICE.md.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.