DaFeiYu QQ Plugin
An AI chat plugin that connects to QQ private chats and group chats via the OneBot protocol. It supports binding multiple personas, DSH drawing scheduling, cross-machine collaboration and group management tools, and can enrich context and transfer files across devices.
Timeline (last 90 days)
The Big Fat Fish QQ plugin connects to QQ private chats and group chats via the OneBot protocol.
The plugin allows private chats and group chats to use different personas separately, and each binding can independently select the model and host access permissions.
The plugin feeds group members' names, group nicknames, titles and quoted messages into the context, and supports message types such as text, images, voice, files, chat cards and merged forwards integrated into the same message stream.
The plugin offers group management tools such as group announcements, muting, recalls and group albums, and can view group chat status.
The plugin receives messages in real time, merges and understands consecutive interjections, and gives a unified response after the other party has finished speaking.
At-completion marks workspaces, canvases and files; the underlying layer uses unique identifiers for locating, and display names serve only as interface hints.
Cross-canvas references are read-only snapshots, preventing a change in one place from affecting another.
In cross-machine collaboration, after selecting a paired host, both sides communicate over an encrypted connection; the remote side decides which directories may be read and which tasks and tools may run, with the permission boundary controlled by the accessed party.
When referencing materials across machines, a host prefix replaces the local workspace reference syntax, while workspace names and relative paths can be reused.
Remote canvases can run in the background, and the local machine can view progress, add instructions or stop them; remote questions are brought back to the local machine and continue in the original canvas, and final replies are returned in their original order with their source preserved.
2 more timeline events
All Facts (12)
The Big Fat Fish QQ plugin connects to QQ private chats and group chats via the OneBot protocol.
50%UnverifiedThe plugin allows private chats and group chats to use different personas separately, and each binding can independently select the model and host access permissions.
50%UnverifiedThe plugin feeds group members' names, group nicknames, titles and quoted messages into the context, and supports message types such as text, images, voice, files, chat cards and merged forwards integrated into the same message stream.
50%UnverifiedThe plugin offers group management tools such as group announcements, muting, recalls and group albums, and can view group chat status.
50%UnverifiedThe plugin receives messages in real time, merges and understands consecutive interjections, and gives a unified response after the other party has finished speaking.
50%UnverifiedAt-completion marks workspaces, canvases and files; the underlying layer uses unique identifiers for locating, and display names serve only as interface hints.
50%UnverifiedCross-canvas references are read-only snapshots, preventing a change in one place from affecting another.
50%UnverifiedIn cross-machine collaboration, after selecting a paired host, both sides communicate over an encrypted connection; the remote side decides which directories may be read and which tasks and tools may run, with the permission boundary controlled by the accessed party.
50%UnverifiedWhen referencing materials across machines, a host prefix replaces the local workspace reference syntax, while workspace names and relative paths can be reused.
50%UnverifiedRemote canvases can run in the background, and the local machine can view progress, add instructions or stop them; remote questions are brought back to the local machine and continue in the original canvas, and final replies are returned in their original order with their source preserved.
50%UnverifiedAuthorized remote plugins can be called directly, with requests executed on the target host without creating a new remote canvas.
50%UnverifiedThe multi-host trust scenario with multiple devices paired pairwise, authorized separately, and each capable of acting as primary or backup controller has not yet been tested in the demo video.
50%