Social MCP Connector
A remote MCP server that provides tools to query live Meta (Facebook+Instagram) and TikTok organic social data, such as follower counts, insights, recent posts, and aggregated overviews.
README
Social MCP Connector
A remote Model Context Protocol server that exposes your Meta (Facebook + Instagram) and TikTok organic social data as MCP tools, so you can ask Claude Desktop or ChatGPT things like "how did my Instagram account grow this week?" or "list my latest TikToks with their view counts."
It hits the live Meta Graph API and TikTok Display API using access tokens you provide as environment variables — there is no mock data.
- Transport: Streamable HTTP + SSE via the Vercel
mcp-handleradapter (@vercel/mcp-adapter). - Stack: Next.js (App Router) + TypeScript, deploys to Vercel.
- Endpoint: a single route handler at
app/api/[transport]/route.ts.
Connector URLs
Once deployed to https://<your-app>.vercel.app:
| Transport | URL |
|---|---|
| Streamable HTTP (recommended) | https://<your-app>.vercel.app/api/mcp |
| SSE (legacy clients) | https://<your-app>.vercel.app/api/sse |
SSE resumability across serverless invocations needs Redis. It is optional — leave
REDIS_URLunset and the Streamable HTTP path (/api/mcp) works fully without it.
TikTok OAuth helper endpoints
| Endpoint | Purpose |
|---|---|
GET /api/tiktok/auth |
Redirects to TikTok to authorize and generate a token. |
GET /api/tiktok/callback |
Shows the resulting TIKTOK_ACCESS_TOKEN to copy into Vercel. |
Start here to get your token: https://<your-app>.vercel.app/api/tiktok/auth
Tools
See tools.md for full parameter and return details.
| Tool | What it does |
|---|---|
instagram_account_overview |
IG followers, media count, name, bio, profile picture |
instagram_account_insights |
IG reach, profile views, follower growth over a period |
instagram_recent_media |
Recent IG posts with per-post likes / comments / reach |
facebook_page_insights |
FB Page impressions, post engagements, fan count |
tiktok_user_info |
TikTok followers, following, likes, video count |
tiktok_recent_videos |
Recent TikTok videos with views / likes / comments / shares |
social_overview |
Aggregated headline numbers across IG + FB + TikTok |
Environment variables
Copy .env.example to .env.local for local dev, or set these in
Vercel → Project → Settings → Environment Variables.
| Variable | Required? | Description |
|---|---|---|
META_ACCESS_TOKEN |
for Meta/IG tools | Long-lived Meta access token with the permissions below. |
META_IG_USER_ID |
for IG tools | Instagram Business/Creator account id (numeric, not the @handle). |
META_PAGE_ID |
for FB tool | Facebook Page id linked to the IG account. |
META_APP_ID |
optional | App id — only for token debugging / appsecret_proof. |
META_APP_SECRET |
optional | App secret — only for token debugging / appsecret_proof. |
TIKTOK_ACCESS_TOKEN |
for TikTok tools | User access token from TikTok Login Kit OAuth. |
TIKTOK_CLIENT_KEY |
for OAuth helper | TikTok app client key. Used by /api/tiktok/auth to mint the access token. |
TIKTOK_CLIENT_SECRET |
for OAuth helper | TikTok app client secret. Used by /api/tiktok/callback for the token exchange. |
TIKTOK_REDIRECT_URI |
optional | Override the OAuth redirect URI. Defaults to ${origin}/api/tiktok/callback. Must match the URI registered in the TikTok app. |
MCP_AUTH_TOKEN |
optional | If set, the MCP endpoint requires Authorization: Bearer <value>. If unset, the endpoint is open. |
REDIS_URL |
optional | Enables SSE resumability. Not needed for Streamable HTTP. |
The build never reads tokens.
npm run buildmakes no API calls and passes with no env vars set; tokens are only used at request time.
Meta prerequisites
- An Instagram Business or Creator account, linked to a Facebook Page.
- A Meta app with these permissions granted on the token:
instagram_basic,instagram_manage_insights,pages_read_engagement,pages_show_list,read_insights(andbusiness_managementif using a Business portfolio). - A long-lived token (short-lived tokens expire in ~1 hour). Exchange via:
GET https://graph.facebook.com/v21.0/oauth/access_token?grant_type=fb_exchange_token&client_id=<APP_ID>&client_secret=<APP_SECRET>&fb_exchange_token=<SHORT_LIVED_TOKEN> - You usually only need
META_ACCESS_TOKEN. The server auto-discovers the rest: it callsGET /me/accountsto find the Page you manage (and its Page access token, which Page insights require) and the linked Instagram Business account id.META_PAGE_ID/META_IG_USER_IDare optional overrides — if set and valid they win; if missing or wrong they are auto-resolved. Each tool's output reports which id was used and itssource(env vs discovered). To set them explicitly:GET https://graph.facebook.com/v21.0/me/accounts?access_token=<TOKEN>(Page id), thenGET https://graph.facebook.com/v21.0/<PAGE_ID>?fields=instagram_business_account&access_token=<TOKEN>(IG user id).
TikTok prerequisites
The TikTok portal only gives you a Client Key + Client Secret — the access token must come from the Login Kit OAuth flow. This app includes a helper to do that for you:
- A TikTok developer app with Login Kit and the Display API enabled.
- Set
TIKTOK_CLIENT_KEYandTIKTOK_CLIENT_SECRETin Vercel (+ optionallyTIKTOK_REDIRECT_URI) and deploy. - In your TikTok app, register the Redirect URI:
https://social-mcp-connector.vercel.app/api/tiktok/callback(it must match exactly — including https and no trailing slash differences). - Add the scopes:
user.info.basic,user.info.profile,user.info.stats,video.list. - Sandbox mode: add your TikTok account as a target user in the app's sandbox, otherwise authorization fails with a scope/permission error.
- Visit
https://social-mcp-connector.vercel.app/api/tiktok/authin a browser. It redirects you to TikTok to authorize, then the callback page displays youraccess_token,refresh_token,expires_in, andopen_id. - Copy the
access_tokeninto Vercel asTIKTOK_ACCESS_TOKENand redeploy.
The OAuth helper never logs or stores tokens — it only renders them once on the callback page for you to copy.
Deploy to Vercel
- Push this repo to GitHub (already done if you cloned it from
davolu). - In Vercel, Add New → Project and import the repo. Framework preset: Next.js (auto-detected). No build settings to change.
- Add the environment variables above under Settings → Environment Variables (Production + Preview).
- Deploy. Your MCP endpoints are then:
https://<your-app>.vercel.app/api/mcphttps://<your-app>.vercel.app/api/sse
Local dev (optional): npm install then npm run dev (do not commit .env.local).
Add the connector in Claude Desktop
- Open Claude Desktop → Settings → Connectors.
- Click Add custom connector.
- Name:
Social MCP(anything). - URL: paste your Streamable HTTP endpoint:
https://<your-app>.vercel.app/api/mcp - If you set
MCP_AUTH_TOKEN, add an Authorization header with the valueBearer <your-MCP_AUTH_TOKEN>(in Claude Desktop's connector auth field, or the OAuth/headers section). IfMCP_AUTH_TOKENis unset, leave auth empty. - Save / Connect. The seven tools above will appear and you can start asking questions like "Give me a social_overview" or "Show my recent Instagram media."
ChatGPT: use Settings → Connectors → Create / Add custom connector and paste the same
/api/mcpURL (plus the bearer token if configured).
Robustness notes
- Every tool checks for its required token first and returns a clear message
(e.g.
❌ TIKTOK_ACCESS_TOKEN is not set...) rather than throwing. - A small fetch helper (
lib/fetcher.ts) adds a 20s timeout and normalises Meta and TikTok API error shapes (expired token, missing permission, rate limit) into readable messages. - A single tool/API failure never crashes the server;
social_overviewreturns partial results with per-platform errors inline.
Security
- Tokens are read from server-side env vars only and are never sent to the browser or included in responses.
.env*is git-ignored (except.env.example); no secrets are committed.- Set
MCP_AUTH_TOKENto require a shared-secret bearer token on the endpoint.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.