Appa Tools documentation for MCP Studio, including setup, guides, concepts, and API-related reference content.

Skip to main content

Create an MCP server

The MCP Studio wizard walks you through source selection, tool selection, analytics configuration, review, and deployment. Use this guide to create a server that gives AI agents the right context from your trusted sources.

AudienceMCP creators, technical writers, developers, support teams
PrerequisitesAppa account access, source URLs such as docs, sites, repos, or PDFs, and an MCP client for testing

Add sources​

Paste URLs for the content you want your MCP to reference. Supported source types:

  • Documentation sites: For example, https://docs.docker.com.
  • GitHub repos: Public repos or private repos. Private repos require GitHub sign-in.
  • Websites: Blog posts, wikis, help centers, or knowledge bases.
  • PDFs: Uploaded or linked PDF content that should become searchable context.
  • MCP endpoints: Existing MCP servers that you want to federate into your context workflow.
Large docs sites

If a docs site is very large, the wizard will recommend either:

  • Providing a more specific URL, such as https://docs.docker.com/engine/.
  • Connecting the docs GitHub repo for broader coverage.

Website sources are indexed up to 5,000 pages per source. Very large sites can still benefit from narrower URLs for higher precision and faster indexing.

Choose tools​

Select between 3 and 10 tools for your MCP server. Each tool gives AI agents a specific capability. See the full Tools Reference for details on each one.

Configure analytics​

Choose which Core metrics to track. You can also enable Action or Predictive analytics tiers for deeper quality and impact reporting.

Review the server​

Preview your server name, sources, and selected tools. Make any final adjustments.

Deploy the server​

Click Deploy. MCP Studio creates your MCP server and returns the following connection details:

  • A Server URL for streamable HTTP JSON-RPC (default)
  • A Legacy SSE endpoint for compatibility-only clients
  • A JSON config snippet ready to paste into your AI tool

Your server starts indexing sources in the background as durable jobs. You do not need to keep the dashboard open. The server is usable immediately — it uses live fetching as a fallback until indexing completes.