Modules

Kickside Content Renderers — isolated, reusable content-viewer web components. <wc-resource-view> host picks + mounts a viewer (<wc-render-markdown>/<wc-render-pdf>/<wc-render-image>) by content-type; each viewer is its own lazy bundle so heavy deps never enter a host. Used by the wiki, uploads, inbox, anywhere a resource needs inline rendering.

Kickside Hub — module discovery and lifecycle management. Ships the kickside-hub System surface plus a standalone onboarding component that recommends real catalog modules and reviews selected installs sequentially through Keeper.

Universal external-identity linking — map a provider account (discord, slack, telegram, …) to a Kickside user via short-lived dashboard-issued codes, so any bot resolves "who is this external sender" provider-neutrally.

Kickside security — scope and policy administration surface for app-owned security scopes.

Kickside UI — platform navigation API plus admin UI surfaces. Ships the kickside-appearance web component (a System tab) that edits navigation layout, categories, and per-surface overrides backed by kickside.ui settings.

Module specifications and documentation

Knowledge system interface contracts

Kickside realtime — a centralized contract for pushing topics to connected UIs. broadcast() fans out to every client via the wippy/relay central hub; notify() targets one user's hub. Any module (nav, inbox, …) publishes through one seam so all UI reacts live.

Kickside Data Sync automation kind — source-to-destination synchronization built on the shared automation engine.

Telegram bot webhook handler with registry-based command routing

MCP (Model Context Protocol) server for Wippy — exposes tools to LLM clients over stdio

Knowledge base system with embedding and query support

Typed executable Blocks for headless and visual Flow composition.

Userspace components for user management and integrations

Recursive web-based code editor with LSP support

Kickside knowledge bases — KB registry, document/graph KB schema and data-access stores, ingestion/injection workers, concept-graph enrichment, and a user-facing CRUD API.

Optional visual authoring, versioning, and run lifecycle for published Workflows composed from canonical Blocks and native Dataflow nodes.

Render a PDF — from a URL or a binary blob — into per-page PNG, JPEG, or WEBP images

Web research tools for AI agents - search, read pages, and extract links

Model Context Protocol client