MCP knowledge server
Articles, workshop slides and the workshop catalog by Mike Bild as an MCP server for AI assistants, read-only and without login.
MCP knowledge server
All articles of this website, the workshop slides from slides.mikebild.dev and the public workshop catalog are available as an MCP server (Model Context Protocol). Your AI assistant searches the content directly, reads individual articles in full and cites the original URL. Articles and slides exist in German and English.
- Endpoint:
https://worker.mikebild.dev/workshop-online-api/knowledge/mcp - Transport: Streamable HTTP, read-only, no login, no API key
- Source: the same files this website is built from, as knowledge.json
Setup#
Claude Code:
claude mcp add --transport http mikebild-knowledge https://worker.mikebild.dev/workshop-online-api/knowledge/mcp
Codex (~/.codex/config.toml):
[mcp_servers.mikebild-knowledge]
url = "https://worker.mikebild.dev/workshop-online-api/knowledge/mcp"
Clients with a JSON configuration, for example Cursor:
{
"mcpServers": {
"mikebild-knowledge": {
"url": "https://worker.mikebild.dev/workshop-online-api/knowledge/mcp"
}
}
}
Tools#
| Tool | What it returns |
|---|---|
search_articles |
full-text search across all articles, hits with canonical URL, summary and text excerpt |
get_article |
one article as Markdown with date, topics and a link to the audio version |
list_articles |
articles by topic or year, plus all topics with their counts |
list_decks |
workshop slides with PDF, video (when narrated), exercise repository, related articles and slide count |
get_deck_outline |
outline of a slide deck: every slide with number, title, section and link, plus PDF, exercise repository and articles |
list_workshops |
publicly listed workshops from workshops.mikebild.dev |
Rules#
- The content here remains the original. Cite the URL you receive when you use something from it.
- The server changes nothing and stores no requests beyond the technical access log.
- The endpoint is limited to a few requests per second. For bulk queries use llms-full.txt or knowledge.json.
- Further machine-readable outputs: llms.txt, RSS feed, Blogcast.
Questions or requests for further tools: Contact.