Why three platform layers

Lingchu Bot organizes platform support around three layers: platform, protocol, and implementation. This separation keeps each layer’s behavior, capabilities, and configuration concerns isolated instead of mixing them into a single page.

Layer model

Layer	Purpose	Examples
Platform	The chat network the bot connects to	QQ
Protocol	The wire/API standard the adapter implements	OneBot V11
Implementation	A concrete server that speaks the protocol	NapCat, default

A platform profile (in platforms/registry.py) declares the known protocols and their priority. Each protocol maps to a NoneBot adapter package. Within a protocol, one or more implementations may extend the default handler set with implementation-specific APIs.

Why three layers

Collapsing platform, protocol, and implementation into one concept loses information that matters at runtime. The same QQ platform can be reached through the OneBot V11 protocol, but the implementation running that protocol determines which APIs are available — NapCat exposes _send_group_notice for group announcements, while a generic OneBot V11 server does not. The three-layer model lets the menu filter commands by the exact implementation detected, rather than advertising commands that would fail.

Available platforms

Platform	Protocols	Status
QQ	OneBot V11	Implemented
Telegram	Telegram Bot API	Implemented

For protocol-side runtime setup, see the connection guides under How-to → Connect.

Adding new platforms

Telegram is available through nonebot-adapter-telegram; see the Telegram connection guide. Discord and other platform profiles remain candidates for future work. See the adapter system reference for the contribution workflow.

How commands are scoped

Command handlers live under handle/<platform>/adapters/<protocol>/<implementation>/. The menu system filters visible commands by the active platform, protocol, and implementation at runtime, so implementation-specific extension APIs (for example, group announcements on NapCat) are hidden unless the current runtime is known to support them.

See the QQ platform reference for protocol- and implementation-specific command references.