Prebuilt modules, or designed logic

BotGhost is the mature pick-and-click platform with 30+ ready modules. BotShade brings a DAG editor, reusable functions, typed persistent variables, and a multi-LLM AI to the same no-code surface — so you design behavior instead of bolting presets together.

A DAG editor with reusable functions

BotShade's custom command builder is a DAG built on a visual node editor with roughly 60 node types. Three nodes — a function definition node, a function call node, and a return node — turn any subgraph into a typed subroutine with arguments, return values, and recursion up to depth 32 and 1000 calls per run. BotGhost's block editor has Loops and Conditions, but its documentation does not describe a comparable concept of argument-typed reusable functions.

Typed persistent variables with atomic RMW

BotShade stores string, number, boolean, object, array, url, money, and time values alongside typed Discord references (user, role, channel, guild), nested up to five levels for shapes like nested arrays of objects. The equation update node wraps updates in row-level locking to prevent lost writes. BotGhost exposes four primitive types — Text, Number, Collection, Object — across Global, Server, Channel, and User scopes, with Free plans capped at 500 variable calls per month.

AI treated as a first-class citizen

BotShade's AI command builder generates a command build plan as an intermediate representation from natural language, then applies it as a reviewable diff with undo. BotGhost's AI module is scoped to ChatGPT and DALL-E, requires a Premium plan plus your own OpenAI API key, and does not document support for Claude or other vendors.

Structured conditions, loops, and error handlers

Condition nodes in BotShade (comparison, chance, channel, role, user, permission, regex) automatically generate and lock the required branches — the else branch, the post-condition continuation, and the matching match branch — and the loop family pairs the loop body and after-loop branch with built-in flow controls for breaking out, continuing, and waiting. A dedicated error-handler subtree exposes the error message as a variable expression, and the recursion guard raises a warning rather than blocking, so recursive calls remain available.

Components V2 and a variable-aware editor

BotShade builds Discord Components V2 messages (Section, Container, MediaGallery, File, Separator) and pairs them with a variable-aware autocomplete that completes server, user, role, and command-option references inline as you type. BotGhost also supports Components V2; the differentiator on the BotShade side is the inline variable-expression completion sitting on top of it.

Per-bot diagnostics with variable provenance

BotShade streams structured logs per bot, labeled by the action that produced them (for example, a custom command blocked from overwriting a protected variable), and tags every variable in the catalog by its origin — user-defined, declared by a plugin's configuration, or installed from the marketplace — to stop marketplace plugins from silently hijacking a name. BotGhost's log surface centers on Real-time Bot Alerts, gated behind Premium Priority Hosting.

Side-by-side comparison

Where BotShade and BotGhost diverge in design, AI, data, and operations.

No-code design

BotShade

BotGhost

Editor model

DAG of nodes and edges on a visual node editor

Block-based actions, conditions, options, and loops

Node / block variety

About 60 node types

12 action categories

Functions (typed subroutines)

function definition / function call / function return nodes, depth 32, 1000 calls per run

Loops and Conditions exist; argument-typed functions are not documented

Error handlers

error-handler subtree with {Error.message}

No structured error-handler primitive in the docs

Pre-save validation

Node, config, and function validation are enforced in the editor

No equivalent pre-save validation layer is documented

AI

BotShade

BotGhost

LLM support

AI command builder generates a command build plan; the LLM layer is pluggable

ChatGPT (GPT-3.5 Turbo through GPT-4) and DALL-E only

Billing model

AI is provided by the platform

Premium plan plus your own OpenAI API key — billed twice

Command generation

Natural language → preview → apply with diff review and undo

Auto-reply chat mode and a respond command, aimed at conversation

Claude / multi-LLM

Architecture abstracts the LLM provider

No mention of Claude or Anthropic in the docs

Persistent data and variables

BotShade

BotGhost

Type system

string, number, boolean, object, array, url, money, time, plus user / role / channel / guild references

Four types: Text, Number, Collection, Object

Scopes

global, server, user, server-member

Global, Server, Channel, User

Nesting

Up to five levels of nesting, including arrays of objects

Nesting depth is not documented

Concurrent writes

Atomic read-modify-write via row-level locking

Locking strategy is not documented

Free-tier limit

Effectively unlimited today (no monthly call cap on Free)

500 variable calls per month on Free, unlimited on Premium

Hosting and operations

BotShade

BotGhost

Bot token ownership

You bring your own Discord application and token, and run it as your bot

Mostly a shared platform model; personalization is layered on top

Bot state control

Start, stop, and restart bots directly from the dashboard with a single action

Plugin on/off toggles rather than process-level state

Diagnostics

Per-bot structured logs keyed by task name, with date-range filtering

Real-time alerts gated behind Premium Priority Hosting

Extensibility

BotShade

BotGhost

Marketplace

Plugin marketplace where each plugin declares the settings it needs, users fill them in through a setup panel, and references are restricted to those declared settings.

Fixed catalog of built-in modules

Variable provenance

Every variable is tagged with its origin: user-defined, declared by configuration, or supplied by the marketplace

No provenance model documented

You design the program, not just configure presets

Typed functions, structured conditions and loops, atomic persistent variables, and an AI that writes the graph for you add up to something closer to a small visual language than a settings page.

Multi-LLM by design, no double billing

The AI builder treats the model as a swappable backend, and you do not pay both a platform fee and your own OpenAI bill to use it.

A deep catalog of ready modules

30+ prebuilt modules covering moderation, reaction roles, leveling, music, tickets, and more — fast to enable when you want the standard set without designing anything.

Familiar block editor

The action-and-condition block model is approachable for teams that have used MEE6, Carl-bot, or earlier BotGhost projects.

When to choose BotShade

You want to build real logic — functions, loops, typed state, AI-generated graphs — and run the bot as your own Discord application with structured diagnostics.

When to choose BotGhost

You want a broad library of prebuilt modules you can switch on quickly, and the built-in command primitives cover what you need without custom logic.

Coming from BotGhost?

Most BotGhost commands map cleanly onto BotShade nodes — actions become nodes, conditions become condition subtrees, and variables move into the typed catalog. We can help you map your existing modules and rebuild the parts that depend on a function or recursion that BotGhost cannot express.

Talk to support →