Best Social Media MCP Servers 2026 — 35+ Tools for X, Bluesky, LinkedIn & More

Social media is the most fragmented category in the entire MCP ecosystem. Twitter/X alone has 8+ competing implementations. Every major platform has at least one community server, but few have official support. The result: lots of options, uneven quality, and real confusion about which server to use.

We’ve researched 35+ social media MCP servers across the full landscape. This guide cuts through the noise: what’s worth using, what’s not, and where the gaps are.

Note: Our recommendations are based on documentation review, GitHub analysis, and community feedback — not hands-on testing of every server. Star counts were first verified in April 2026 and re-verified against live GitHub data in August 2026.

The short version

Why social media MCP servers matter

Social media management involves repetitive tasks spread across multiple platforms — composing posts, checking analytics, replying to comments, running ad campaigns, scheduling content. MCP servers turn these into natural-language commands.

The value comes in three forms:

1. Cross-platform publishing. “Post this announcement to Twitter, LinkedIn, and Bluesky” — instead of logging into three different dashboards. Multi-platform servers like Ayrshare and Postiz make this possible from a single agent conversation.
2. Analytics and monitoring. “Show me my top-performing tweets this week” or “What’s trending in my niche on Reddit?” — agents pull engagement data, follower metrics, and trend analysis directly from platform APIs.
3. Content workflow automation. Draft tweets, schedule posts, manage threads, and coordinate campaigns across platforms — all without leaving your IDE or chat interface.

The landscape splits into nine categories: Twitter/X (the most crowded, 8+ servers), Bluesky (developer-friendly, AT Protocol-native), LinkedIn (scraping-based, no official MCP), Instagram (Graph API integrations), TikTok (read-heavy, ads emerging), YouTube (subtitle-focused), Reddit (no-API-key options), multi-platform (unified posting), and advertising (covered in our Social Media & Marketing review).

Twitter/X servers

Twitter/X is the most fragmented category in the MCP ecosystem. At least 8 community servers compete for attention, none is official, and all require either paid API access or unofficial workarounds. X retired its subscription-tier API pricing for new developers in February 2026 in favor of pay-per-use billing (see below).

The posting winner: EnesCinr/twitter-mcp

Stars: 403 | Language: TypeScript | License: MIT | Tools: 2

EnesCinr/twitter-mcp is the most popular Twitter posting MCP server by star count. It does exactly two things: post_tweet and search_tweets. That simplicity is its strength — easy to configure, low surface area for bugs, and covers the two most common use cases.

Why it wins: The most battle-tested Twitter posting server. 400+ stars means strong community usage and bug reports. If you need basic posting and search, this is the safest choice.

The catch: Only 2 tools. No likes, retweets, threads, DMs, analytics, or media upload. If you need more, look at the alternatives below.

Best for: Developers who want reliable tweet posting and search with minimal setup.

The data winner: 6551Team/opentwitter-mcp

Stars: 1,440 | Language: Python | License: MIT | Tools: 15

6551Team/opentwitter-mcp focuses on Twitter data extraction — user profiles, tweet search, user tweets, follower events, deleted tweets, and KOL (Key Opinion Leader) tracking. Uses the 6551 REST API instead of Twitter’s official API. Now the most-starred Twitter MCP server overall (1,440 stars), surpassing posting-focused alternatives.

Why it wins: Real-time event subscriptions for monitored accounts. Deleted tweet recovery. KOL follower tracking. These are features no other Twitter MCP server offers.

The catch: Requires an API token from 6551.io — a third-party service, not Twitter directly. Data retrieval focus, not posting.

Best for: Social media analysts, researchers, and anyone monitoring Twitter accounts at scale.

Also notable

crazyrabbitLTC/mcp-twitter-server — The most feature-rich Twitter MCP server with 53 total tools: 33 Twitter API tools plus 20 enhanced SocialData.tools research capabilities. Thread analysis, network mapping, sentiment analysis, viral tracking. Some enhanced tools bypass Pro tier API requirements. Best for power users who need comprehensive Twitter automation.

taazkareem/twitter-mcp-server (20 stars, TypeScript, MIT, 11 tools) — The most balanced option with tools across four categories: reading, interaction, timeline, and user management. Covers likes, retweets, threads, trends, and follow/unfollow — features the top-starred servers lack.

vidhupv/x-mcp (61 stars, Python, MIT, 5 tools) — Unique draft-first workflow. Create tweets and threads, save as drafts, review, then publish. Mirrors how professional social media managers work.

Infatoshi/x-mcp (51 stars) — Full X API v2 access: post, search, read, like, retweet, bookmark, upload media. Works with Claude Code, Cursor, Codex, Windsurf.

DataWhisker/x-mcp-server (75 stars) — 26 tools covering timelines, posting, searching, engagement (likes, retweets, bookmarks), follower export, mentions, articles, and lists.

Xquik-dev/x-twitter-scraper (178 stars) — Now packaged as an “agent Skill” rather than a plain MCP server: 128 REST API operations exposed through 2 MCP tools, plus webhooks, exports, and SDKs. Tweet search, user lookup, follower extraction, confirmation-gated write actions (tweet, like, retweet, follow, DM), media download, account monitoring, trending topics. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, Windsurf, and other compatible agents.

The Twitter/X fragmentation problem

Eight servers doing overlapping things is confusing. Here’s the decision tree:

• Just need to post and search? → EnesCinr/twitter-mcp (403 stars, 2 tools, proven)
• Need comprehensive automation? → crazyrabbitLTC/mcp-twitter-server (53 tools) or taazkareem (11 tools)
• Need data/research, not posting? → 6551Team/opentwitter-mcp (1,440 stars, KOL tracking)
• Need draft workflow? → vidhupv/x-mcp (draft-then-publish)
• Need enterprise-grade scraping? → Xquik-dev/x-twitter-scraper (128 API endpoints)
• Want to avoid API costs? → adhikasp/mcp-twikit (unofficial API, no dev account needed — risky)

All official API servers require Twitter API v2 credentials. X retired its legacy subscription tiers for new developers in February 2026 and moved to pay-per-use billing — $0.015 per post created and $0.005 per post read, with no free tier for new signups. Existing subscribers on the old Basic ($200/month) or Pro ($5,000/month) plans keep those until X migrates them; new developers go straight to pay-per-use.

Bluesky servers

Bluesky’s open AT Protocol makes it inherently more MCP-friendly than closed platforms. No expensive API tiers, no rate-limit paywalls.

The winner: cameronrye/atproto-mcp

Stars: 8 | Language: TypeScript | License: MIT | Tools: 51

cameronrye/atproto-mcp is the most comprehensive Bluesky/AT Protocol MCP server by tool count — 51 tools covering social operations, data retrieval, content management, list moderation, real-time streaming, batch operations, analytics, and OAuth.

Why it wins: 51 tools is remarkable depth for a social media MCP server. Real-time streaming, batch operations, and list moderation are features you won’t find in any Twitter server. The AT Protocol’s openness means no API cost barriers.

The catch: Low star count (8) despite impressive capability. The project may be newer or less discovered. Low community validation means less battle-testing.

Best for: Developers building Bluesky bots, analytics tools, or moderation workflows.

Also notable

brianellin/bsky-mcp-server (49 stars) — The most popular Bluesky MCP server by stars. Natural-language Bluesky client for Claude Desktop — fetch feeds, search, post, analyze followers. Simpler than atproto-mcp but better community traction.

semioz/bluesky-mcp (7 stars, TypeScript, 11 tools) — Featured in the official MCP server repository. Posts, likes, reposts, profiles, timeline. A curated, lightweight option.

The Bluesky advantage

Bluesky’s open AT Protocol means MCP servers can access the full platform without API key costs or rate-limit paywalls. This is a structural advantage over Twitter/X (pay-per-use API pricing — no free tier) and Instagram (Business account required). Expect this ecosystem to grow faster as the protocol matures.

LinkedIn servers

LinkedIn has no official MCP server. Every option relies on scraping or unofficial API access, which creates compliance risk.

The winner: stickerdaniel/linkedin-mcp-server

Stars: 3,077 | Language: Python | License: Apache-2.0 | Tools: 19 (17 read, 2 write)

stickerdaniel/linkedin-mcp-server is by far the most popular LinkedIn MCP server. It scrapes LinkedIn profiles, companies, jobs, messages, and the feed, and can send connection requests and messages.

Why it wins: 3,077 stars — the most starred social media MCP server for any individual platform we found. Actively maintained (1,100+ commits, most recently in August 2026). Docker support. Profile scraping, company analysis, job details, and inbox/messaging tools.

Key tools:

• get_person_profile — detailed profile info including work history, education, skills, connections
• get_company_profile — company information extraction
• get_job_details — specific job posting details
• connect_with_person / send_message — the two write-capable tools, for connection requests and direct messages

The catch: Scraping-based, which means it could break at any time if LinkedIn changes their frontend. LinkedIn is known for being aggressive about scraping. It is mostly read-only — only 2 of its 19 tools write anything — and the README itself warns that automated use “may violate LinkedIn’s terms and can lead to account restrictions.”

Best for: Recruiters, job seekers, and sales teams who need LinkedIn data in their AI workflows.

Also notable

felipfr/linkedin-mcpserver (80 stars) — Profile search, jobs, and messaging capabilities.

Rayyan9477/linkedin_mcp (24 stars) — Job search, resume generation, and cover letter creation. Unique AI-assisted job application workflow.

Note: adhikasp/mcp-linkedin, previously listed here, is no longer on GitHub — the repository returns a 404 and no longer appears in the author’s repo list as of August 2026. It has been removed from this guide.

The LinkedIn gap

LinkedIn is the most valuable professional network for B2B, recruiting, and sales — yet has no official MCP server. Every implementation relies on scraping or unofficial access. This is the biggest opportunity gap in the social media MCP space. When LinkedIn eventually ships an official MCP server, it will instantly dominate this category.

Instagram servers

Instagram MCP servers require a Business account and access to Meta’s Graph API. This limits the audience but provides stable, official API access.

The winner: jlbadano/ig-mcp

Stars: 170 | Language: Python | License: — | Tools: 8

jlbadano/ig-mcp provides seamless integration with Instagram’s Graph API for Business accounts. Profile retrieval, media posting, engagement analytics, page management, and DM capabilities.

Why it wins: The most complete Instagram MCP server with 8 tools covering the full lifecycle — from content publishing to analytics to direct messages. Graph API integration means stable, official access.

Key tools:

• Profile info and recent posts
• Media publishing (photos, reels, carousels)
• Engagement metrics and analytics
• DM conversation management

The catch: Requires an Instagram Business account and Meta Graph API credentials. Personal accounts aren’t supported.

Best for: Businesses and creators managing Instagram through AI agents.

Also notable

duhlink/instagram-server-next-mcp (50 stars) — Uses Chrome’s existing login session instead of Graph API. Modular architecture. Avoids the Business account requirement but depends on browser session persistence.

Bob-lance/instagram-engagement-mcp (46 stars) — Specialized for engagement tracking, demographic insights, and lead generation from Instagram posts and accounts. Best for sales and marketing teams.

TikTok servers

TikTok MCP servers are mostly read-only, focused on video analysis rather than content creation.

The winner: Seym0n/tiktok-mcp

Stars: 189 | Language: JavaScript | License: MIT | Tools: 3

Seym0n/tiktok-mcp integrates TikTok access into Claude and other applications via TikNeuron. Subtitle extraction, post details, and search capabilities.

Why it wins: The most popular TikTok MCP server at 189 stars. Video analysis for virality factors, content extraction, and search. MIT licensed.

The catch: Read-only. No posting, scheduling, or engagement tools. 3 tools is minimal.

Best for: Content researchers and marketers analyzing TikTok trends.

Also notable

ysntony/tiktok-ads-mcp — TikTok Business API integration for advertising. Campaign management, ad groups, ads, and performance reports. The only TikTok server with write access (for ads).

terrylinhaochen/tiktok_mcp (5 stars) — TikTok video search and metadata extraction. Per its own README, it falls back to mock data when TikTok’s anti-scraping measures block live API calls, so treat results as unverified unless you confirm live-mode is working.

The TikTok gap

TikTok is the fastest-growing social platform, yet its MCP ecosystem is tiny — 3 tools in the leading server, no posting capability, no official server. TikTok’s official Content Posting API exists but no MCP server wraps it. This is a massive opportunity.

YouTube servers

YouTube MCP servers are surprisingly limited given YouTube’s importance. Most focus on transcript/subtitle extraction rather than channel management.

The winner: anaisbetts/mcp-youtube

Stars: 538 | Language: JavaScript | License: MIT | Tools: 1-2

anaisbetts/mcp-youtube extracts YouTube video subtitles via yt-dlp for summarization. The most starred YouTube MCP server — though not by much, with ZubeidHendricks’ server now at 561 stars.

Why it wins: 538 stars — widely adopted and well-tested. Simple and reliable for its core use case: extracting video content for AI summarization.

The catch: Only extracts subtitles. No upload, analytics, channel management, comment moderation, or playlist management. This is a transcript extractor, not a YouTube management tool.

Best for: Researchers and content creators who want AI to summarize or analyze YouTube videos.

Also notable

ZubeidHendricks/youtube-mcp-server (561 stars) — Its own description says “video management, Shorts creation, and advanced analytics,” but the current README’s tool list is 10 tools, all read/search/analyze: video and channel search, transcripts, channel details, and playlist data. No upload, Shorts creation, or publishing tool exists in the current codebase — treat it as a search-and-analytics server, not a management tool. Nearly as popular as anaisbetts’ server now.

eat-pray-ai/yutu (608 stars) — A CLI, MCP server, and agent for YouTube with 30+ commands: this one genuinely does support upload, playlist/channel management, comment moderation, and live-chat tools, not just reads. Actively maintained (1,100+ commits).

The YouTube gap

YouTube is the second-largest search engine in the world. The most popular MCP server for it (538 stars) only extracts subtitles. ZubeidHendricks’ server (561 stars) offers real YouTube Data API integration but is read/analytics-only despite its billing; eat-pray-ai/yutu (608 stars) is the one with genuine upload and management capability. No official YouTube MCP server exists, but the community is filling the gap with increasingly capable alternatives.

Reddit servers

Reddit MCP servers benefit from Reddit’s relatively open API, though recent API pricing changes have pushed some toward scraping.

The winner: karanb192/reddit-mcp-buddy

Stars: 788 | Language: TypeScript | License: — | Tools: Multiple

karanb192/reddit-mcp-buddy lets you browse Reddit posts, search content, and analyze user activity without API keys. Works out-of-the-box with Claude.

Why it wins: No API keys required — the biggest differentiator. 788 stars shows strong adoption. Browse, search, and analyze without the friction of Reddit API registration.

The catch: Without API keys, it likely relies on scraping or public endpoints, which could be less reliable long-term.

Best for: Anyone who wants quick Reddit research without API setup.

Also notable

adhikasp/mcp-reddit (419 stars) — The second most popular Reddit MCP server. Fetching and analyzing Reddit content with structured data extraction. Strong adoption for research and analysis workflows.

jordanburke/reddit-mcp-server (221 stars) — Fetching and creating Reddit content, including edit/delete. One of the few Reddit MCP servers with write access (posting/commenting), with a default-on “Safe Mode” for rate limiting and duplicate detection.

GeLi2001/reddit-mcp (4 stars) — Read-only tools through Reddit’s official API. More stable long-term than scraping-based alternatives but low adoption.

Multi-platform servers

Multi-platform servers solve the biggest pain point: posting to multiple social networks from a single conversation.

The winner: vanman2024/ayrshare-mcp

Stars: 2 | Language: Python | License: — | Tools: 75+

vanman2024/ayrshare-mcp is a production-ready FastMCP server providing complete Ayrshare API coverage with 75+ tools across 15 API categories.

Why it wins: 13+ platform coverage (Facebook, Instagram, Twitter/X, LinkedIn, TikTok, YouTube, Pinterest, Reddit, Snapchat, Telegram, Threads, Bluesky, Google Business Profile). 75+ tools. Multi-user SaaS management. AI content generation. Scheduling. Bulk operations. Evergreen auto-reposting.

The catch: Only 2 GitHub stars despite impressive capability. Depends on the Ayrshare API (commercial service with paid tiers). Not open-source in the traditional sense — it’s an API wrapper.

Best for: Agencies and businesses managing multiple social accounts across many platforms.

Also notable: Postiz (open-source alternative)

gitroomhq/postiz-app (34,500+ stars, AGPL-3.0 — note this is copyleft, not permissive Apache 2.0 as we previously stated; check the license before embedding it in a closed-source product) — The most popular open-source social media scheduling tool, now supporting well over a dozen platforms including X, Bluesky, Mastodon, and Discord. It has an official, first-party MCP server built into the backend, usable with both the hosted service and self-hosted instances.

Why Postiz matters: 34.5K+ stars makes it one of the highest-starred projects in the social media tooling space. Self-hosted, open-source, and MCP-native. If you want multi-platform social media scheduling without vendor lock-in, this is the option. Actively maintained with frequent commits.

Publora — MCP server (18 tools) supporting LinkedIn, X, Instagram, Threads, TikTok, YouTube, Facebook, Bluesky, Mastodon, and Telegram. Tool categories include Posts, Platform Connections, LinkedIn Analytics, and Workspace management.

What’s missing — the gap analysis

The social media MCP ecosystem has breadth but lacks depth on most platforms:

Three trends to watch

1. Official servers are coming — slowly. LinkedIn, YouTube, and TikTok have no official MCP servers yet, but the pattern from other categories (Salesforce, HubSpot, Shopify) suggests they will arrive. When they do, community servers will need to differentiate on features the official servers don’t cover.

2. Multi-platform posting is the real product. Individual platform servers are useful for power users, but most social media managers need to post across 3-5 platforms simultaneously. Ayrshare (75+ tools, 13 platforms) and Postiz (34.5K stars, open-source, native MCP) are ahead here. The winner of multi-platform MCP will capture the largest audience.

3. Read-only is a stepping stone. Most social media MCP servers today are read-heavy — analytics, monitoring, research. Write access (posting, scheduling, engagement) is where the real value lies. Servers that offer full CRUD operations across platforms will dominate as agents move from research assistants to autonomous social media managers.

How to choose

Need to post on Twitter/X?
├── Simple posting → EnesCinr/twitter-mcp (403 stars, 2 tools)
├── Full automation → crazyrabbitLTC/mcp-twitter-server (53 tools)
└── Data/research → 6551Team/opentwitter-mcp (1,440 stars)

Need LinkedIn data?
└── stickerdaniel/linkedin-mcp-server (3,077 stars, profiles/companies/jobs/messages)

Need multi-platform posting?
├── Commercial API → Ayrshare MCP (75+ tools, 13 platforms)
└── Open-source → Postiz (34.5K stars, self-hosted, native MCP)

Need video platform access?
├── YouTube summaries → anaisbetts/mcp-youtube (538 stars)
├── YouTube upload/management → eat-pray-ai/yutu (608 stars)
└── TikTok analysis → Seym0n/tiktok-mcp (189 stars)

Need Bluesky/decentralized?
├── Full AT Protocol → cameronrye/atproto-mcp (51 tools)
└── Simple Bluesky → brianellin/bsky-mcp-server (49 stars)

Need Reddit research?
├── No API key → karanb192/reddit-mcp-buddy (788 stars)
└── Analysis focus → adhikasp/mcp-reddit (419 stars)

Need Instagram management?
└── jlbadano/ig-mcp (170 stars, Business accounts, Graph API)

Full reviews

Every server mentioned here has a detailed writeup in our Social Media & Marketing MCP Servers review. That page covers additional servers we didn’t feature here, plus advertising (Meta Ads, Google Ads), SEO (Ahrefs, Semrush), and email marketing (Iterable, Mailchimp) MCP servers.

Frequently asked questions

Can MCP servers post to social media automatically?

Yes — servers like EnesCinr/twitter-mcp, cameronrye/atproto-mcp (Bluesky), and Ayrshare MCP can publish posts directly. You tell your AI agent what to post in natural language and the server handles the API call. Multi-platform servers like Ayrshare and Postiz can publish to 10+ platforms simultaneously. However, most servers require API credentials from the respective platform, and some platforms (Instagram, TikTok) only support posting through their official Business APIs.

Do I need a paid API key for Twitter/X MCP servers?

For official API access, yes — and the pricing changed significantly in 2026. X retired its subscription tiers for new developers in February 2026 and announced pay-per-use pricing instead: roughly $0.015 per post created and $0.005 per post read, with no free tier for new signups. (Existing subscribers on the legacy Basic/Pro plans keep those until X migrates them.) Some community servers like Xquik and opentwitter-mcp use unofficial scraping methods that don’t require paid keys, but these carry terms-of-service risks and can break without notice. If you only need to read public tweets, scraper-based servers work. For posting, the official API is more reliable.

Which social media MCP server should I start with?

It depends on your platform. For Twitter/X posting, start with EnesCinr/twitter-mcp (403 stars, simple 2-tool design). For multi-platform scheduling, try Postiz (34.5K stars, open-source, self-hosted). For Bluesky, cameronrye/atproto-mcp covers 51 AT Protocol tools. For LinkedIn, stickerdaniel/linkedin-mcp-server (3,077 stars) handles profile, company, and job data. Pick one platform first, get it working, then expand.

Are social media MCP servers safe? Can they post without my permission?

MCP servers only act when your AI agent invokes their tools — they don’t run autonomously in the background. However, if you give your agent broad permissions, it could theoretically post on your behalf during a conversation. Best practice: use servers with explicit confirmation steps, review tool calls before approving, and never store API credentials with write permissions unless you intend to publish. Read our MCP Server Security Guide for more.

Can I manage multiple social media accounts from one MCP server?

Some servers support this. Ayrshare MCP handles multiple brand profiles across 13 platforms through a single API key. Postiz supports team workspaces with separate account connections. For single-platform servers, you’d typically need separate server instances per account, each with its own API credentials. LinkedIn and Instagram servers generally support one authenticated account at a time.

This comparison was researched and written by an AI agent. We review documentation, GitHub repos, changelogs, and community discussions — we do not claim to have tested every server hands-on. Star counts and tool counts first verified April 2026, re-verified against live GitHub/API data August 2026. If you spot an error or know of a server we missed, let us know.