The Model Context Protocol (MCP) is an open standard, introduced by Anthropic in late 2024, that lets AI assistants connect to external tools, data, and workflows through one shared interface. It works like a universal plug: once a platform exposes an MCP endpoint, any MCP-compatible client (Claude, Cursor, VS Code, GitHub Copilot, ChatGPT, Windsurf, and others) can read from it without a custom integration.

The social.plus MCP server is a hosted endpoint that exposes the social.plus knowledge base to any MCP-compatible AI assistant. It serves the documentation, API reference, OpenAPI specs, SDK guides, and UIKit reference in real time, so your AI tool can answer questions, generate code, and reason about social.plus integrations using the same source material our engineering team writes against.

social.plus officially documents setup for Claude Desktop, Claude Web, VS Code (with GitHub Copilot, version 1.99 or later), and Cursor. Because the Model Context Protocol is an open standard, any other MCP-compatible client (Windsurf, Claude Code, ChatGPT through the Apps SDK, agentic frameworks like LangChain) should also connect. New clients are added by their vendors, not by us.

Paste https://learn.social.plus/mcp into your AI client's MCP configuration. The setup is officially documented for Claude Desktop, Claude Web, VS Code (with GitHub Copilot, version 1.99 or later), and Cursor, each with a short JSON snippet or in-app form. Restart the client and the social.plus tools appear in your assistant. Full per-client guides at learn.social.plus/ai-mcp/overview. No keys, no local server, no auth.

No. Connecting to the social.plus MCP server requires no account, no API key, and no configuration beyond pasting the URL into your AI client. You are free to explore social.plus capabilities through your assistant immediately. When you are ready to build and ship social.plus in your own application, that production setup is a separate step. Talk to our team to get started.

The social.plus MCP server is read-only and exposes only the public knowledge base: documentation, API reference, OpenAPI specs, SDK and UIKit guides. It never touches customer data, application traffic, end-user information, or live network state. Your data remains exactly where it always has, inside your own social.plus environment.

Use the MCP to build, debug, and explore social.plus integrations faster. Your AI assistant can scaffold a chat feature, generate UIKit code, look up the correct API endpoint for a moderation rule, debug a webhook, or compare SDK methods, all without leaving your editor. It supports every stage of work: evaluating social.plus, integrating it for the first time, or maintaining a production deployment.

The social.plus MCP server is for in-app community infrastructure, not social media management. It exposes the APIs, SDKs, and UIKit you use to embed chat, feeds, communities, and live streaming inside your own application. Social media MCP servers, by contrast, automate posting on platforms like X, Instagram, or TikTok. social.plus is the layer beneath your product, not a destination network.

The documentation requires you to stop, search, copy, and adapt snippets into your AI tool by hand. The MCP server eliminates that hop. Your AI assistant queries the same documentation, API reference, OpenAPI specs, and SDK guides directly in the middle of your coding task, then produces the answer or generates the code in place. Less context-switching, and every reference your assistant uses is the current one.

The Model Context Protocol itself is an open standard, maintained by Anthropic with reference implementations and SDKs published on modelcontextprotocol.io. The social.plus MCP server is operated by social.plus and provides curated, current platform context to any compatible client. The protocol is open; the server is hosted by us so the content stays accurate.

Yes. The social.plus MCP server exposes the full product catalog, including Chat, Social, Live Stream, Communities, Moderation, and UIKit, with the capabilities, limits, and configuration of each documented inline. Describe your use case to your AI assistant, whether that is a fitness app with group challenges, a marketplace with buyer-seller messaging, or a creator platform with live shopping, and it maps your scenario to the right combination of social.plus SDKs and surfaces the methods you should start with.

Yes. The social.plus MCP server exposes the full SDK reference for every supported platform, including Swift for iOS, Kotlin for Android, React Native, Flutter, JavaScript, and TypeScript. When you ask your AI assistant to add chat, feeds, live streaming, or moderation, it selects the correct SDK for your stack, generates the right initialization sequence, and follows the established integration patterns from the docs. No copy-pasting between language-specific docs.

Yes. That is one of the main reasons the MCP server exists. With your AI assistant connected, you describe the feature in plain language, the assistant generates the initialization code, scaffolds the UIKit components, and points you at the social.plus sandbox to test against. A working prototype in a single session is realistic, and the same loop continues into production once you decide to commit.

Yes. With the MCP server connected, your AI assistant can read your existing social.plus integration code, cross-reference it against the current API and SDK reference, and flag deprecated calls, missing webhook handlers, or misconfigured parameters. The same workflow extends to adding capabilities later, whether that is adding live streaming on top of an existing chat integration, layering moderation on a community feed, or migrating from an older SDK version.

The social.plus MCP server is purpose-built for this. It exposes the social.plus SDKs, APIs, and UIKit components for chat (1:1, group, channel), communities (activity feeds, posts, comments, reactions), and live streaming (broadcasting, viewer playback, embedded chat), all designed for embedding inside your own application. Ask your AI assistant to add any of these features, and it pulls the correct integration code and configuration from the live social.plus knowledge base, no matter which client you connect from.

Yes. The social.plus MCP server is the community MCP for in-app community infrastructure. It exposes the SDKs, APIs, and UIKit components for activity feeds, posts, comments, reactions, communities, group memberships, and roles. When connected to your AI assistant, it takes you from "add a community space to this screen" to working integration code without manually navigating the docs. social.plus communities power consumer apps at scale, including Noom, Smart Fit, and Harley-Davidson.

Yes. The social.plus MCP server is the social features MCP for in-app development. It exposes the SDKs and UIKit components for the building blocks of social experiences: activity feeds, user profiles, posts, comments, reactions, likes, shares, follows, mentions, and notifications. With the MCP connected, your AI assistant generates the right initialization, the correct field shapes, and the UI components that match your design system, all drawn from the live social.plus reference.

Yes. The social.plus MCP server doubles as a live streaming MCP for AI-assisted development. It exposes the social.plus Live Stream SDK and UIKit reference, covering broadcasting, viewer playback, low-latency ingest, embedded chat overlay, reactions, and recording. Ask your AI assistant to add a live commerce stream, a live event broadcast, or a live coaching session, and it generates the integration code against the current SDK methods, ready to test against the social.plus sandbox.

The social.plus MCP server is for adding social-network-style features to your own application, not for connecting AI to an external social network. Searches for 'social network MCP' today return two distinct categories. The first is MCP servers that integrate AI with platforms like Discord, Bluesky, or X for posting and analytics. The second is the social.plus MCP, which exposes the in-app community infrastructure (feeds, follows, profiles, communities, and chat) needed to embed those experiences inside your own product, with full data ownership and brand control.

Yes. The social.plus MCP server includes the moderation and trust and safety surfaces of the platform: AI-assisted content review, profanity and image filters, role-based permissions, automated rules, and the admin Console workflows for human review. Ask your AI assistant to configure moderation for your use case, whether that is a teen-focused app, a regulated industry, or a global community with multilingual content, and it surfaces the correct API endpoints, rule structures, and Console steps from the social.plus knowledge base.

Yes. The social.plus MCP server is the activity feed MCP for AI-assisted development. It exposes the Social SDK and UIKit components for activity feeds, posts, comments, reactions, custom post types, and ranking logic. Ask your AI assistant to add a curated feed, a community wall, or a personalized timeline to your app, and it generates the integration against the current SDK methods, ready to test against the social.plus sandbox.

Yes. The social.plus MCP server covers the Stories and Clips features alongside the rest of the platform, including upload, ephemeral expiry, replies, view tracking, and the UIKit components used to render the story tray and full-screen viewer. Ask your AI assistant to add stories to your home screen, and it generates the integration code, the lifecycle handlers, and the moderation hooks in one pass.

Yes. The social.plus MCP server includes the live commerce and social commerce surfaces of the platform, where product cards, checkout actions, and shoppable interactions attach directly to the live stream overlay, activity feed, or stories layer. Ask your AI assistant to add a shoppable live stream or a feed post with a buy action, and it generates the integration against the current SDK and UIKit components.

Yes. The social.plus MCP server exposes the full feature set, including polls (single-vote, multi-vote, real-time results), events (scheduling, RSVPs, attendance), groups (topic-based spaces, member roles, permissions), group chat (multi-participant real-time conversations), and custom posts (structured content types beyond standard posts). When you describe the engagement pattern you want, your AI assistant pulls the correct API endpoints and UIKit components for it.

Yes. The social.plus MCP server includes implementation patterns documented for the verticals social.plus serves: retail (live commerce, post-purchase communities), fitness (challenges, group goals, community feeds), travel (group itineraries, traveler chat), sports and betting (live chat overlays, real-time reactions), health and wellness (gated communities, accountability groups), fintech (verified profiles, secure DMs), media and news (comments, reactions, audience polls), edtech, and gaming. Ask your AI assistant for the pattern that fits your vertical, and it returns the right combination of SDKs and configuration.

The social.plus MCP server endpoint is https://learn.social.plus/mcp. Paste this URL into your AI client's MCP configuration to connect. One endpoint serves everything: documentation, API reference, SDK guides, UIKit, and the eight skill domains (admin, chat, server, setup-validator, social, troubleshoot, uikit, video). The transport is SSE (Server-Sent Events), and no API key or account is required.

The social.plus MCP server exposes two tools and eight skill domains. The tools are search_social_plus for natural-language search across the knowledge base, and query_docs_filesystem_social_plus for sandboxed, read-only filesystem queries that support rg, grep, find, cat, jq, awk, sed, and other Unix-style commands against the docs and OpenAPI specs (stateless, with a 30KB output cap per call). The skill domains are admin, chat, server, setup-validator, social, troubleshoot, uikit, and video, organized so your AI assistant scopes its lookups to the right area without scanning the entire knowledge base.

Open the Claude Desktop config file at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS, or %APPDATA%\Claude\claude_desktop_config.json on Windows (create the file if it does not exist), then add this entry:

{
  "mcpServers": {
    "social-plus": {
      "url": "https://learn.social.plus/mcp",
      "transport": "sse"
    }
  }
}

Quit and reopen Claude Desktop, and social-plus appears under MCP servers in the toolbar. No API key, no local server, no further configuration.

VS Code requires version 1.99 or later and the GitHub Copilot extension. Open the user settings JSON (Cmd+Shift+P on macOS, Ctrl+Shift+P on Windows or Linux, then run Preferences: Open User Settings (JSON)) and add this configuration at the top level:

{
  "mcp": {
    "servers": {
      "social-plus": {
        "type": "sse",
        "url": "https://learn.social.plus/mcp"
      }
    }
  }
}

Open Copilot Chat with Cmd+I, switch to Agent mode, and run @social-plus to confirm the tool list appears.

Open Cursor settings with Cmd+, on macOS or Ctrl+, on Windows or Linux, navigate to Features → MCP, select Add new MCP server, and enter:

• Name: social-plus
• Type: SSE
• URL: https://learn.social.plus/mcp

The server displays as Enabled in the MCP list once saved. Alternatively, add the same configuration directly to .cursor/mcp.json in your project root.

Verification depends on the client. In Claude Desktop, look for social-plus listed under MCP servers in the toolbar after the restart. In VS Code, open Copilot Chat in Agent mode and type @social-plus to see the tool list. In Cursor, open Settings → Features → MCP and confirm social-plus shows as Enabled. If any of these are missing, the configuration was not picked up or the client was not restarted.

The most common causes are: forgetting to restart the AI client after editing the config (required in every client), running VS Code older than version 1.99 (the minimum for MCP), the GitHub Copilot extension not installed in VS Code, a typo in the URL (must be exactly https://learn.social.plus/mcp), or pasting the JSON in the wrong location inside the config file. The social.plus MCP server itself includes a setup-validator skill that your AI assistant can use to walk through the diagnosis once any MCP connection is live.