Skip to content

PostMCP MCP

Remote MCP server (stdio or Streamable HTTP, API key or OAuth) - the official PostMCP AI server that connects social media publishing pipelines directly into AI assistants, IDE workflows and web environments. Fifteen tools cover workspaces, connected accounts and their token health, brand kits, the post queue, pre-flight checks, create/schedule/publish/retry/delete, and image generation, across LinkedIn, X (Twitter), Facebook, Instagram, Threads, Bluesky and YouTube Shorts.

Server type: Remote (Streamable HTTP) or stdio (npx)
Auth: API key (env, query param or header) or OAuth 2.0 PKCE (RFC 9728)
Endpoint: self-hosted HTTP mode; stdio via npx -y @postmcpai/server
Tools: 15 (accounts, brand kits, post queue, pre-flight, publishing, image generation, batching)
Pricing: PostMCP AI workspace plans with credit-based post costs
Category: Marketing
Built by: PostMCP AI (MIT, github.com/postmcp/postmcp-mcp-server)

Why This Matters for Operators

Publishing is where agents usually stop being useful: they draft the post and then a human copies it into a scheduler, per platform, per profile. PostMCP's server gives an agent the whole pipeline - connected accounts, brand kits, the queue and the publish action itself.

The tooling is designed for production use rather than demos. A preflight_post dry run checks character limits, unconnected profiles, missing media and credit cost before anything publishes. A multicall tool runs up to 20 operations in one request with tool names validated before anything executes, so a typo cannot leave half a batch written. Every account's token health is queryable, so an agent can detect an expired connection before a scheduled post silently fails.

Tools & Capabilities

Tool Purpose
get_user_info Authenticated user: plan, credit balance, AI tokens, active workspace
list_workspaces Every workspace the user belongs to, with roles and connected platforms
get_connected_accounts Connected social profiles with the profileId needed to target them
get_account_health Connections whose token expired or is close to it
list_brandings Brand kits: tone, audience, keywords, style images
list_posts / get_post Post queue with per-profile delivery status, live URLs and per-profile errors
preflight_post Dry run: character limits, unconnected profiles, missing media, credit cost
create_post Draft, schedule or immediately publish; each profile becomes its own post
publish_post_now Publish immediately; also retries failed posts, skipping delivered profiles
update_post / reschedule_post Edit content, targets, schedule or media; move a post to a new slot
reset_stuck_post / delete_post Release a post stuck mid-publish; cancel a scheduled or failed post
generate_image Generate a post image and return its hosted URL for mediaUrl
multicall Run up to 20 of the above in one ordered request

Installation

npx -y @postmcpai/server

For HTTP mode, set the PORT environment variable to launch the Streamable HTTP server. The API key is read from POSTMCPAI_API_KEY, a URL query parameter, or an x-api-key header.

Configuration

{
  "mcpServers": {
    "postmcpai": {
      "command": "npx",
      "args": ["-y", "@postmcpai/server"],
      "env": {
        "POSTMCPAI_API_KEY": "pmcp_sec_your_api_key"
      }
    }
  }
}

Claude.ai connects over OAuth 2.0 with PKCE (RFC 9728) dynamic client registration. Every tool accepts an optional workspaceId; the API key carries its own workspace by default.

Business Relevance

  • Marketing operators run multi-platform publishing from one agent conversation instead of five dashboards
  • Agencies hold client brand kits in the workspace and pre-flight every post before it schedules
  • Founders get token-health visibility so a lapsed connection never eats a launch announcement
  • Product teams batch create and reschedule posts with the multicall tool

Integration with CorpusIQ

PostMCP is the publishing rail; CorpusIQ is the measurement rail. An agent composing both can draft a launch post in PostMCP's queue while CorpusIQ connectors pull the numbers that make the post credible - Stripe revenue, GA4 signup growth, Shopify sales - and then pre-flight the post against per-platform limits before scheduling. After publication, CorpusIQ reads the traffic and revenue the campaign produced, and the agent decides what to retry or reschedule. The loop is draft, verify with real business data, publish, then measure.

Limitations

  • Brand new listing with no track record yet
  • Publishing itself requires a PostMCP AI workspace and per-post credits
  • YouTube Shorts is the video surface; no long-form publishing yet
  • Self-hosted HTTP mode is the remote path; the npm package is the standard install
Contact Terms Privacy Pricing