Post to X/Twitter from Claude — No API Keys, No Setup

Post to X/Twitter from Claude -- No API Keys, No Setup

Posting to X from an AI assistant should be simple. In practice, it is anything but. You need X API credentials, you need to handle OAuth with PKCE, you need to manage rate limits that change based on your API tier, and as of February 2026, you cannot even reply to most posts programmatically unless the author mentioned you first.

FeedSquad's MCP server eliminates all of this. You add one configuration line. You connect your X account with a click. You tell Claude or ChatGPT what to post. Done.

This guide covers the setup, the workflow, and how FeedSquad compares to the DIY alternatives.

The X API Problem

Before we get into the solution, here is why posting to X programmatically is harder than it looks in 2026.

API tier confusion. X offers Free, Basic ($200/month), Pro ($5,000/month), and Enterprise tiers. Each has different rate limits for posting, reading, and searching. The Free tier allows 500 posts per month but only 1 app and limited read access. The limits are not always obvious until you hit them.

PKCE authentication. X requires OAuth 2.0 with PKCE (Proof Key for Code Exchange) for user-context authentication. This is more secure than basic OAuth but significantly more complex to implement. You need to generate code verifiers, compute SHA-256 code challenges, and store state securely.

Reply restrictions. Since February 2026, X restricts API-based replies on all self-serve tiers. You can only reply if the original author mentioned you or quoted your post. This limitation does not affect original posts, but it is a surprise for developers who expect full API access.

Rate limit management. Post creation is limited per 15-minute window and per 24-hour period, varying by tier. Exceeding limits results in temporary lockouts that can disrupt scheduled content.

FeedSquad handles all of this. You do not need an X developer account, API keys, or knowledge of PKCE. The managed service deals with authentication, rate limiting, and API quirks.

Setting Up FeedSquad for X

Step 1: Add the MCP Server

For Claude Code, add this to your configuration:

{
  "mcpServers": {
    "feedsquad": {
      "type": "http",
      "url": "https://feedsquad.com/api/mcp"
    }
  }
}

For ChatGPT, add FeedSquad as a connected action from the tool settings.

Step 2: Authenticate

The first time you use a FeedSquad tool, you will be prompted to log in or create an account. This is a one-time browser-based OAuth flow.

Step 3: Connect X

Tell your AI assistant:

"Connect my X account"

FeedSquad generates an OAuth URL with PKCE already configured. Click the link, authorize in your browser, and you are connected. FeedSquad stores your tokens securely and handles refresh automatically.

Step 4: Start Posting

"Create a post for X about our new feature launch. Keep it under 280 characters."

FeedSquad creates the draft, checks content quality, and shows you a preview. Approve it and choose to publish now or schedule for later.

FeedSquad vs DIY X MCP Servers

Several open-source X MCP servers exist. Here is how they compare to FeedSquad.

OpenTweet

OpenTweet offers 12 tools covering the tweet lifecycle. It is purpose-built for X and requires an OpenTweet account starting at $11.99/month. It handles X API credentials on your behalf but is X-only.

DataWhisker

An open-source MCP server with 16 tools for X (reading timelines, posting, searching, engagement). You need your own X API credentials, and you run the server locally.

Infatoshi x-mcp

Another open-source option for posting, searching, and engaging on X. Requires your own API keys and local setup.

How FeedSquad Differs

| Aspect | FeedSquad | OpenTweet | DIY Servers | |--------|-----------|-----------|-------------| | API keys needed | No | No (managed) | Yes | | Setup | 1 config line | Account + config | Install + configure + API keys | | OAuth/PKCE | Fully managed | Managed | Handle yourself | | Platforms | X + LinkedIn + Threads | X only | X only | | Content calendar | Built-in | No | No | | Campaign management | Yes | No | No | | Content quality checks | Yes | No | No | | Voice matching | Yes | No | No | | Rich UI in Claude | Interactive cards | No | No | | Scheduling | Built-in | Basic | No |

The single-platform servers are good for developers who want direct X API access. FeedSquad is built for founders and marketers who want to manage their social presence across platforms from their AI assistant.

Cross-Platform Workflow

FeedSquad's strongest advantage is cross-platform publishing. Here is what that looks like in practice:

You: "I just launched a new pricing page. Create a post about it for LinkedIn, X, and Threads. LinkedIn should be detailed and professional. X should be short and punchy. Threads should be casual."

Claude creates three drafts, each adapted for the platform. LinkedIn gets a multi-paragraph post with a hook. X gets a concise post under 280 characters. Threads gets a conversational take.

You: "Schedule all three. LinkedIn tomorrow at 9am, X at 11am, Threads at 2pm."

All three are scheduled. Your content calendar now shows the full cross-platform plan.

This workflow replaces:

• Three separate tools or browser tabs
• Copy-pasting and manually reformatting
• A spreadsheet to track what goes where and when
• Individual authentication with each platform's API

Tips for Posting to X via MCP

Respect the character limit. Tell your AI assistant the constraint: "Keep this under 280 characters." FeedSquad's content health checks will also flag posts that exceed the limit.

Be direct. X audiences scan fast. The best performing posts get to the point immediately. When prompting your AI, say "Make this punchy and direct" rather than asking for explanations or narratives.

Adapt, do not copy. If you are cross-posting from LinkedIn, do not just truncate. Ask your AI to reimagine the core idea for X's format and audience. A 1500-character LinkedIn post should become a sharp observation, not a compressed summary.

Use campaigns for consistency. If you are building an audience on X, consistency matters more than individual post quality. Use FeedSquad's campaign feature to plan a week or month of content in one sitting.

Know the limitations. FeedSquad publishes original posts to X. Due to X's February 2026 API restrictions, automated replies are limited to conversations where you have been mentioned. Plan your content strategy around original posts and quote tweets.

What About Threads Too?

If you are already using FeedSquad for X, adding Threads takes one command:

"Connect my Threads account"

Same OAuth flow, same simple process. Then you can create and publish to X and Threads (and LinkedIn) from a single conversation. See our complete Threads MCP guide for details.

Pricing

FeedSquad's free tier includes 10 posts per month across all platforms. For founders posting regularly on X, the unlimited plan at 29 euros per month removes the cap and adds campaigns, scheduling, and analytics.

See the full pricing page for plan comparison.

Frequently Asked Questions

Do I need my own X API credentials?

No. FeedSquad handles all X API integration, including OAuth 2.0 with PKCE, token management, and rate limiting. You connect your X account through a browser-based authorization flow.

Which X API tier does FeedSquad use?

FeedSquad uses its own X API access. You do not need any X developer account or API subscription.

Can I post threads (multi-tweet posts)?

FeedSquad currently supports single-post publishing to X. Thread support (multi-post sequences) is in development.

What about X's reply restrictions?

FeedSquad focuses on original post creation, not automated replies. The February 2026 API reply restrictions on self-serve tiers do not affect original post publishing.

Can I schedule X posts for specific times?

Yes. You can schedule posts for any future date and time. FeedSquad's content calendar shows all scheduled posts across platforms and prevents timing conflicts.

Will my X posts sound like AI?

Not if you use FeedSquad's voice training feature. It learns your writing style from samples you provide and applies your tone, vocabulary, and patterns to generated content. Combined with content health checks that catch AI-detectable patterns, your posts will sound like you. Learn more about repurposing content across platforms.