Connect Your AI Coding Assistant

Before integrating StreamLayer, connect your AI coding assistant to the documentation MCP server. It lets your assistant look up the current SDK guides and platform integration skills while working in your app.

Choose your assistant below, add the connection, then verify it. You only need to configure each assistant once for the scope you choose (your user account or this project).

The server provides two read-only tools: search finds pages and fetch reads a result. It does not expose your StreamLayer account, control StreamLayer Studio, or run your app. You still need the SDK setup and testing steps for your platform.

Choose your assistant

• Cursor
• Codex
• Claude Code
• GitHub Copilot in VS Code
• Other MCP-compatible tools

Use the instructions for the assistant running your coding session. Their configuration formats differ even though they connect to the same server. If a configuration file already exists, merge this server entry into it; do not replace your other servers.

Cursor

Add this to .cursor/mcp.json in your project, or ~/.cursor/mcp.json to make it available across your projects:

{
  "mcpServers": {
    "streamlayer-docs": {
      "url": "https://docs.streamlayer.io/mcp"
    }
  }
}

Open a Cursor Agent session, enable the server if prompted, and run the verification prompt below. See Cursor's MCP documentation for configuration and server controls.

Codex

With the Codex CLI installed, run:

codex mcp add streamlayer-docs --url https://docs.streamlayer.io/mcp

Alternatively, add this entry to ~/.codex/config.toml:

[mcp_servers.streamlayer-docs]
url = "https://docs.streamlayer.io/mcp"

Start a new Codex session and run the verification prompt below. For the Codex app or IDE extension, use its MCP settings or supported configuration file on the machine running the session. See Codex's MCP documentation.

Claude Code

With Claude Code installed, run this from your project directory:

claude mcp add --transport http --scope project streamlayer-docs https://docs.streamlayer.io/mcp

This adds the server to the project's .mcp.json. Use --scope user instead if you want the connection across your projects. An equivalent project configuration is:

{
  "mcpServers": {
    "streamlayer-docs": {
      "type": "http",
      "url": "https://docs.streamlayer.io/mcp"
    }
  }
}

In Claude Code, use /mcp to check the connection and accept the project server prompt if shown. Then run the verification prompt below. These steps are for Claude Code; other Claude apps have their own connector settings. See Claude Code's MCP documentation.

GitHub Copilot in VS Code

Open the Command Palette, run MCP: Add Server, choose HTTP, and enter the URL and server name from the table above. Choose workspace scope for this project or user scope for your projects.

For a workspace connection, the equivalent .vscode/mcp.json is:

{
  "servers": {
    "streamlayer-docs": {
      "type": "http",
      "url": "https://docs.streamlayer.io/mcp"
    }
  }
}

Use MCP: List Servers to check the connection. Start a Copilot agent session with the server's tools enabled and run the verification prompt below. See VS Code's MCP configuration reference for session-specific behavior; Copilot in other editors or the CLI may use different configuration.

Other MCP-compatible tools

In your tool's MCP settings, add a remote HTTP server named streamlayer-docs with the URL above. No StreamLayer credentials are needed. Use an agent mode that supports MCP tools, then run the same verification prompt. Follow your client's documentation for its configuration format.

Fallback for clients that only support local stdio

Prefer the direct HTTP connection when available. If your client supports only local stdio servers, a local bridge such as mcp-remote can connect it to the remote endpoint. This fallback requires Node.js and npx on the machine running the assistant.

For clients that use the mcpServers JSON format:

{
  "mcpServers": {
    "streamlayer-docs": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://docs.streamlayer.io/mcp"]
    }
  }
}

Use your client's supported configuration location and package approval process. The bridge runs locally; the documentation server remains remote.

Verify documentation access

Replace the bracketed platform with your target, then paste this into your assistant:

Use the streamlayer-docs MCP to search for the StreamLayer integration guide
for [my SDK platform]. Fetch the matching page.
Return its title, source URL, SDK/package name, and required setup steps.
Do not change code yet. If the page is unavailable or the docs conflict,
report that instead of guessing.

Confirm the assistant actually uses search and fetch, returns source URLs on docs.streamlayer.io, and selects your SDK platform. A plausible answer without tool calls does not verify the connection. Successful search and fetch confirm documentation access, not a working app integration.

Next, follow Integrate with an AI Agent or return to your platform's getting-started page for a ready-to-use integration prompt. Find your platform in the Introduction.

Troubleshooting

If your tool cannot use MCP, give it the relevant guide or the documentation index, and ask it to cite the pages it reads. You can also copy a documentation page into its context.