/serve
Start the wiki web server — browsable Wikipedia-style UI with live research, split-pane editor, AI assist, and themes. Manual trigger only: /wiki-serve, 'start wiki server', 'browse the wiki'.
Claude Code
$ npx -y skills add Oshayr/LLM-Wiki --skill serve --agent claude-code
Codex
$ npx skills add Oshayr/LLM-Wiki --skill serve -a codex -y
Cursor
$ npx skills add Oshayr/LLM-Wiki --skill serve -a cursor -y
Gemini CLI
$ npx skills add Oshayr/LLM-Wiki --skill serve -a gemini-cli -y
GitHub Copilot
$ npx skills add Oshayr/LLM-Wiki --skill serve -a github-copilot -y
OpenCode
$ npx skills add Oshayr/LLM-Wiki --skill serve -a opencode -y
Windsurf
$ npx skills add Oshayr/LLM-Wiki --skill serve -a windsurf -y
SKILL.md
serve.SKILL.mdname: wiki-serve
description: "Start the wiki web server — browsable Wikipedia-style UI with live research, split-pane editor, AI assist, and themes. Manual trigger only: /wiki-serve, 'start wiki server', 'browse the wiki'."
Wiki Serve
Launch a local web server that presents the wiki as a browsable, Wikipedia-style website with live background research, integrated chat, and a split-pane editor with AI assist.
Resolve `.wiki/` from plugin install scope. Auto-create if missing.
Arguments
- **`/wiki-serve`** — start the server (default port 8420)
- **`/wiki-serve stop`** — stop the running server
Startup
The FastAPI web server **only runs when explicitly requested**. There are two ways to start it:
Skill-Based Launch
1. Run the `/wiki-serve` skill from Claude Code (recommended):
/wiki-serve
2. The server automatically:
- Installs dependencies if needed
- Handles page indexing, research workers, and file watching
- Opens your browser to localhost:8420
Manual Launch
Dependencies are installed automatically by the plugin's `SessionStart` hook. To start the server manually from the command line:
python path/to/skills/serve/scripts/server.py --wiki-dir .wiki/ --port 8420
Replace `path/to` with the full path to your LLM-Wiki installation (typically `.claude/plugins/llm-wiki`).
If dependencies are missing, install manually: `pip install -r requirements.txt`
Features
- **Page rendering** — markdown → HTML with wiki-link transformation, red-link detection
- **Live research** — click any red link to trigger background research (stub-first: summary in 30s, full upgrade async)
- **Search** — full-text search with autocomplete
- **Split-pane editor** — markdown + live preview, markdown toolbar, AI assist bar
- **Page type templates** — concept, brainstorming, idea, memory, status, rules, config, skill, plus custom templates from `.wiki/templates/`
- **Chat** — wiki-aware AI assistant in bottom-right panel
- **Research queue** — manage, cancel, reorder research tasks
- **Knowledge graph** — interactive page relationship visualization
- **Three themes** — light, dark, terminal
- **Export** — HTML and markdown download
Stop
Kill the server process on port 8420. Either:
- `/wiki-serve stop`
- The server auto-stops when the session ends (daemon process)
Read more
name: wiki-serve description: "Start the wiki web server — browsable Wikipedia-style UI with live research, split-pane editor, AI assist, and themes. Manual trigger only: /wiki-serve, 'start wiki server', 'browse the wiki'."
Wiki Serve
Launch a local web server that presents the wiki as a browsable, Wikipedia-style website with live background research, integrated chat, and a split-pane editor with AI assist.
Resolve `.wiki/` from plugin install scope. Auto-create if missing.
Arguments
- **`/wiki-serve`** — start the server (default port 8420)
- **`/wiki-serve stop`** — stop the running server
Startup
The FastAPI web server **only runs when explicitly requested**. There are two ways to start it:
Skill-Based Launch
1. Run the `/wiki-serve` skill from Claude Code (recommended):
/wiki-serve
2. The server automatically:
- Installs dependencies if needed
- Handles page indexing, research workers, and file watching
- Opens your browser to localhost:8420
Manual Launch
Dependencies are installed automatically by the plugin's `SessionStart` hook. To start the server manually from the command line:
python path/to/skills/serve/scripts/server.py --wiki-dir .wiki/ --port 8420
Replace `path/to` with the full path to your LLM-Wiki installation (typically `.claude/plugins/llm-wiki`).
If dependencies are missing, install manually: `pip install -r requirements.txt`
Features
- **Page rendering** — markdown → HTML with wiki-link transformation, red-link detection
- **Live research** — click any red link to trigger background research (stub-first: summary in 30s, full upgrade async)
- **Search** — full-text search with autocomplete
- **Split-pane editor** — markdown + live preview, markdown toolbar, AI assist bar
- **Page type templates** — concept, brainstorming, idea, memory, status, rules, config, skill, plus custom templates from `.wiki/templates/`
- **Chat** — wiki-aware AI assistant in bottom-right panel
- **Research queue** — manage, cancel, reorder research tasks
- **Knowledge graph** — interactive page relationship visualization
- **Three themes** — light, dark, terminal
- **Export** — HTML and markdown download
Stop
Kill the server process on port 8420. Either:
- `/wiki-serve stop`
- The server auto-stops when the session ends (daemon process)
An autonomous knowledge base that grows as you work. LLM Wiki is a Claude Code plugin that captures research, ideas, and decisions into an interlinked wiki with semantic search, automatic research, and a Wikipedia-style web UI.
More skills from other plugins.
codebase-wiki
How to work in a Codebase Wiki project (the `codebase-wiki` starter pack): an agent-authored, source-grounded wiki of the…
academic-paper-reviewe…
Simulates academic peer review, evaluating papers across Originality, Methodology, Results, and Writing to provide Major/Minor…
Questions about serve.
- Which coding agents can use serve?
- serve installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.

