8 live
Automate

Automate your socials with Claude Code

Claude Code adds Antwork with one command and authorizes with a slash command. Nothing leaves the terminal, and there is no API key at any point.

Platforms8API keys0To startfree
One command
$ claude mcp add antwork --transport http https://api.antwork.io/mcp

Needs Claude Pro, Max, Team, or an API key.

Connect it

  1. 01Authorize with /mcpType /mcp inside Claude Code and pick antwork. It opens an OAuth consent screen in your browser; approving it makes the connection live and the token refreshes itself.
  2. 02Connect the social accountAsk Claude Code to connect your account and it returns a connect link per platform. Authorize once and Antwork holds and refreshes that token for you.
  3. 03Add the skill (optional)The open-source Antwork skill teaches the agent the right tool order — pull context, draft, then schedule or publish as a second explicit step.

Server details

Server URL
https://api.antwork.io/mcp
Server name
antwork

The server speaks Streamable HTTP and authenticates with OAuth 2.1, using dynamic client registration and PKCE. Claude Code handles the token exchange for you — all you see is a browser consent screen.

If it does not work

The three things that go wrong most often when connecting Claude Code:

  • Antwork's tools do not appear — restart Claude Code after installing. Most clients only read their MCP configuration at startup.
  • The consent screen keeps reappearing — the session expired or the grant was revoked. Approve it once more; grants last 30 days and refresh silently while you keep using them.
  • The tools answer but report no accounts — the workspace exists, but no social account is connected to it yet. Connect one in the Antwork web app first.

Your agent already writes. Let it post.

One command, one consent screen, and every connected account is something you can just ask for.

Connect your AI