What is RBX Debugger
A debugger for live Roblox games. Remote calls, HTTP traffic, the instance tree and decompiled source, read from your browser.
RBX Debugger attaches to a running Roblox game. The interface is a web page. The capture happens in your executor, on your machine.
The tabs#
| Tab | What it shows |
|---|---|
| Web Traffic | HTTP the game makes, with replay and response overrides |
| Spy | FireServer, InvokeServer and inbound OnClientEvent, arguments decoded |
| Explorer | The live instance tree, properties, and a 3D view of a part |
| Inspect | getgc, the registry, environments, instance dumps |
| Tools | Fire prompts and signals, saveinstance, fast flags, hashing |
| Code | Decompiled source for any script in the game |
| Remotes and Modules | What a scan found, with fire and require helpers |
Settings holds the UI zoom and the local-only MCP server. The RBX Editor is its own page at rbx.lol/editor and works with no game attached.
What you run#
The page
rbx.lol/debugger is the whole interface. Nothing to install.
The client
One loadstring in your executor. It hooks __namecall, captures HTTP, walks
the instance tree and serialises arguments, inside your own game session.
The relay
A browser tab cannot accept an inbound connection, so a relay pairs the page and the client by room code and forwards frames between them. It stores nothing. See Privacy and security for what crosses it, and for how to point the client at your own bridge instead.
Requirements#
| Requirement | Why |
|---|---|
An executor with WebSocket.connect | How the client talks to the relay |
hookmetamethod or equivalent | Capturing __namecall for remote traffic |
request / http_request | Capturing and replaying HTTP |
Most modern executors have all three. A missing piece disables one feature.
Without request, HTTP capture is unavailable and remote spying still works.