Platform Core
Unified Capability Directory
Organize AI chat, AI PPT, AI image, AI video, and agent capabilities through one shared registry.
The shared chat surface for multi-model conversations, advisor entry, web search, and enterprise context.
Acts as the shared execution surface for advisor and agent flows.Preferred text route for DeepSeek-family traffic.Preferred text route for OpenAI-family traffic.Available only when explicitly configured or explicitly selected.
Build on the current PPT tooling and ppt-master work to support platform-level preview, export, and templating.
Preview is available before login; export stays behind authentication.Already live as a first public toolsite entry.Supports preview before login and export after login.Ready to expand toward enterprise templates and brand systems.
Keep the current image assistant workspace while elevating it into a platform image capability with references and enterprise workflows.
Image generation defaults to the async assistant queue for stability.The existing image workspace remains reusable.
Extend the current video workspace into a platform surface for generation, editing, and distribution-adjacent entry points.
Text-to-video and image-to-video execute through the official MiniMax Hailuo video API.MiniMax is the default video model; RunningHub remains a workflow fallback when configured.Keeps the current dashboard/video entry.
Adds one entry for soundtrack, voice cloning, and speech synthesis powered by the MiniMax audio runtime.
MiniMax powers AI music, voice clone, and async speech synthesis inside the shared media workspace.Ships as a dedicated public tool page.Shares the same media task adapter as image and video.
Organize agents, plugins, MCP services, and workflows with a registry-first approach backed by existing runtimes.
Preferred text route for DeepSeek-family traffic.Preferred text route for OpenAI-family traffic.Available only when explicitly configured or explicitly selected.