Read Voidreg with your AI client

Every published post and project is available to agents. Connect over MCP to search and read the knowledge base, or fetch plain Markdown. Nothing here needs an account or a token.

Connect over MCP

The server speaks Streamable HTTP and is read only. It never exposes drafts, and it treats page content as data.

Server URL Streamable HTTP, no token
https://www.voidreg.xyz/mcp
Claude Code Run once in a terminal.
claude mcp add --transport http voidreg https://www.voidreg.xyz/mcp
Cursor and other JSON clients Add to the mcpServers section of your MCP settings.
{
  "mcpServers": {
    "voidreg": { "url": "https://www.voidreg.xyz/mcp" }
  }
}
Codex Add to ~/.codex/config.toml.
[mcp_servers.voidreg]
url = "https://www.voidreg.xyz/mcp"

What an agent can do

search_content
Search posts and projects by title, summary or tag.
list_articles
Browse published posts, newest first, by tag or project.
get_article
Read one post as Markdown with its canonical link.
list_projects
Browse projects by category, tag or search.
get_project
Read one project as Markdown, with the posts that follow it.
list_topics
See how posts are grouped inside each project.

Posts and projects are also exposed as resources: voidreg://blog/{slug} and voidreg://project/{slug}.

Without MCP

.md
Add it to any post or project address, or ask the page for text/markdown.
llms.txt
A map of the site: every project, and posts shared out fairly between them.
llms/<project>.txt
Every post in one project, grouped by topic.
llms-full.txt
The text of the site in one document, the newest of each project first.
llms/<project>-full.txt
The full text of one project.
rss.xml
The 50 newest posts as a feed.
sitemap.xml
Every public page, with last-modified dates.