Back to AI Agents

MCP Setup Guide

Connect AdaptlyPost to your AI assistant in under a minute.

Prerequisites

Recommended

Setup for Claude Code & Cursor

Add this to your MCP settings. For Claude Code: <code>~/.claude/settings.json</code>. For Cursor: <code>.cursor/mcp.json</code>.

MCP config

{
  "mcpServers": {
    "adaptlypost": {
      "type": "http",
      "url": "https://mcp.adaptlypost.com/mcp",
      "headers": {
        "Authorization": "Bearer adaptly_your_api_key_here"
      }
    }
  }
}

Get your API key from the API Tokens page. Keys start with adaptly_.

Claude Code CLI

claude mcp add --transport http --scope project --header "Authorization: Bearer adaptly_your_api_key_here" -- adaptlypost https://mcp.adaptlypost.com/mcp

Or use the Claude Code CLI one-liner:

Setup for Claude Desktop

Claude Desktop requires the <code>mcp-remote</code> bridge for HTTP MCP servers. Add this to your <code>claude_desktop_config.json</code>:

claude_desktop_config.json

{
  "mcpServers": {
    "adaptlypost": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.adaptlypost.com/mcp",
        "--header",
        "Authorization: Bearer adaptly_your_api_key_here"
      ]
    }
  }
}

Setup for other MCP clients

Any MCP client that supports HTTP transport can connect directly.

Endpoint

https://mcp.adaptlypost.com/mcp

Auth header

Authorization: Bearer adaptly_your_api_key_here

Transport

HTTP (Streamable HTTP)

For clients that only support stdio transport, use the <code>mcp-remote</code> bridge as shown in the Claude Desktop section.

Alternative: Agent Skill

For Claude Code, Cursor, Windsurf, and Codex you can also install AdaptlyPost as an agent skill β€” no MCP configuration needed.

npx skills add adaptlypost/agent

Installs a SKILL.md and CLI script that teaches your agent the AdaptlyPost API. View on GitHub

Tool Reference

The MCP server exposes 12 tools. Your AI discovers these automatically β€” just ask in plain language.

ToolDescriptionKey params
list_accountsList all connected social media accounts with IDs, platforms, and usernames.None
upload_mediaUpload media from URLs or base64 data. Returns mediaUrls for create_post.urls?, files?
get_upload_urlsGet presigned upload URLs for direct file uploads to storage.files (fileName + mimeType)
create_postCreate and schedule a post to one or more platforms.text, platforms, contentType, timezone, scheduledAt?, mediaUrls?, platformTexts?, saveAsDraft?, tiktokConfigs?, youtubeConfigs?, pinterestConfigs?
get_postGet full details of a single post by ID.id
list_postsList posts with optional filters by platform, status, or date range.statuses?, platforms?, startDate?, endDate?, sortOrder?, limit?, offset?
update_postUpdate a scheduled or draft post's caption, schedule, media, or platform configs.id, text?, platforms?, contentType?, scheduledAt?, mediaUrls?, platformTexts?
delete_postDelete a scheduled or draft post.id
publish_draftPublish a draft post immediately or schedule it for later.id, timezone, scheduledAt?
list_post_resultsCheck per-platform posting results β€” success/failure with error details.id
retry_failed_platformsRetry publishing on platforms that failed.id, platformIds
bulk_schedule_postsSchedule multiple posts at once with shared platforms and connection IDs.platforms, timezone, posts[], connectionIds, platformConfigs?

Managing access

You can view and revoke API tokens anytime from your API Tokens page. Revoking a token immediately disconnects any agent using it.

Ready to connect?

Copy the config above, paste it into your MCP settings, and start posting from your AI.