Use these docs from Claude, Cursor, or any MCP client

This site runs a public, read-only Model Context Protocol server. Connect it once and your AI assistant can search and read every published page — so it answers questions about the Dash Marketing API with current, first-party information instead of guesses.

https://docs.dashmarketing.io/api/mcp

claude mcp add --transport http dash-docs https://docs.dashmarketing.io/api/mcp

{
  "mcpServers": {
    "dash-docs": { "url": "https://docs.dashmarketing.io/api/mcp" }
  }
}

Available tools

• search_docs

  Full-text search across every published page, including code samples and endpoint paths.

• get_page

  Fetch a page's complete content as markdown by its path.

• list_pages

  List all published pages as a tree of titles and paths.

Prefer plain files?

No MCP needed: every page is also available as raw markdown — append .md to any page URL — and /llms.txt indexes everything for tools that discover docs automatically. There's also a “Copy for LLM” action on every page, and the “Ask AI” chat in the corner answers questions with cited sources.

Team access to internal docs

Internal team systems (chatbots, Claude Code, skills) can read internal published pages through the same MCP server by sending an API key — an admin can mint one under Admin → Settings:

claude mcp add --transport http dashdocs https://docs.dashmarketing.io/api/mcp \
  --header "Authorization: Bearer dashdocs_…"

Without a key, the MCP server only exposes published, public pages — never drafts or internal content. Invalid keys are rejected rather than downgraded. Keyed connections also get list_context_files and get_context_file for the team's uploaded reference files (API specs, schemas).

Writing docs from your codebase

A key created with Allow writing drafts also lets an assistant like Claude Code draft docs straight into this site: it can see the full page tree (drafts included), propose where new pages belong, create and update page drafts, and upload screenshots and diagrams to the same image storage the editor uses. Nothing is published automatically — each change comes back with an editor link for a person to review and publish, and edits to existing pages are checkpointed in the page's History.

Write tools: list_page_tree, get_draft, create_page, update_page, upload_image, get_authoring_guide, plus a document_code prompt that walks through the whole workflow. With a shell, images can be uploaded straight from disk:

curl -F "file=@diagram.png" \
  -H "Authorization: Bearer dashdocs_…" https://docs.dashmarketing.io/api/upload