PostMcpAi

What is PostMcpAi?

PostMcpAi is an AI-powered social media management tool that creates, schedules, publishes and measures posts across seven networks: X (Twitter), LinkedIn, Instagram, Facebook, Threads, Bluesky and YouTube Shorts. It pairs a dashboard composer and content calendar with an open-source social media MCP (Model Context Protocol) server, so assistants such as Claude, ChatGPT, Cursor, VS Code, Copilot, Gemini, Windsurf and Zed can draft, schedule and publish through it.

The page states that one API key and one pasted link connect an existing AI client, and that the same server works across MCP clients. A REST API covering all seven networks is also documented for wiring publishing into your own product or automations. Neither the MCP server nor the API is required to use the app.

How to use PostMcpAi?

1. Connect your accounts. Link X, LinkedIn, Instagram and the other supported networks, one click each. Channels are authorized through each platform's official OAuth 2.0 sign-in; PostMcpAi states it never sees or stores social passwords, access tokens are encrypted at rest, and any channel can be revoked from the dashboard.
2. Tell your agent what you want. You can type a plain-English instruction such as "announce our launch on Friday", and the agent writes a version for each network, checking character limits, line breaks, hashtags and tone. Alternatively, write in the composer and let it draft for you.
3. Review and publish. Every publish waits for your approval, so automatic posting is described as never unsupervised. Scheduled posts appear on a visual calendar that can be rearranged, and failed posts can be retried.

For local AI clients, the documented setup is a JSON configuration block pointing at npx -y @postmcpai/server plus an API key from the dashboard; restarting the client makes the publishing tools appear. Browser-based clients such as Claude.ai and ChatGPT connect to a hosted copy of the same server over HTTPS, either as a remote MCP connector or as Custom GPT Actions built from its OpenAPI schema. The server is MIT licensed and published on npm as @postmcpai/server.

Core features of PostMcpAi

AI drafting written per network

The same idea is rewritten for each destination, adapting character limits, line breaks, hashtags and tone rather than cross-posting one block of text.

Visual schedule and calendar

Everything the agent schedules lands on a visual calendar, where posts can be moved or a day cleared before publishing.

Approval gate on every publish

The page states that every publish waits for a yes. This is presented as the one step the agent never skips.

Analytics across seven networks

Open any published post to read views, likes, comments and shares from LinkedIn, X, Facebook, Instagram, Threads, Bluesky and YouTube Shorts, per profile. Connected accounts also report followers, posts and views. Figures are read on request, one credit per read, and never in the background; looking again is free. The same figures can be read by an AI agent over MCP or the API.

Team collaboration

Teammates can be invited into a shared workspace to draft, schedule and approve the agent's posts, with one calendar and one bill. The listed plans differ in seats, workspaces and approvals.

Brand voice stored in your own AI

Tone, audience and go-to phrases are written into your AI's own instructions, so posts routed through PostMcpAi reflect that wording.

Open-source MCP server and REST API

PostMcpAi describes a social media MCP server as a Model Context Protocol server that exposes social publishing as typed tools an assistant can call: create, schedule, publish, retry a failure, read the queue, and check which account tokens are expiring. The site says 16 or 17 such tools are shipped (the page shows both figures) across seven networks, split into read-only and mutating so approval prompts stay meaningful. One binary supports stdio and streamable HTTP transports; social OAuth tokens stay encrypted in the backend and the model sees tool names, arguments and results only.

Companion mobile app in closed beta

An iOS and Android app is in closed beta on TestFlight and Google Play. Applications are approved by hand, and the email used to apply is the one the app checks at sign-in. The app offers the queue and calendar, push notifications, media publishing and AI drafting on the go. The page states that no plan or checkout happens inside the app.

Use cases of PostMcpAi

Announcing a launch across several networks

Give the agent a sentence, a link or a changelog, and it writes a version for each network, checks the limits and schedules it into the calendar.

Keeping a posting queue full

The agent schedules posts at its chosen times and keeps the queue populated, subject to your approval before anything goes live.

Checking post performance without opening seven apps

Read views, likes, comments and shares per post or per connected account from the dashboard, or ask your AI agent the same question over MCP or the API.

Teams sharing publishing duties

A shared workspace lets teammates draft, schedule and approve posts, with roles and plan-dependent teammate limits.

Developers embedding publishing

The REST API covering all seven networks can be used to wire publishing into your own product or automations.

Frequently asked questions about PostMcpAi

Do I need technical skills?

According to the page, none. You connect an AI you already use with one pasted link, or write in the composer, and you can review every draft before it goes out.

Can the AI post automatically?

Yes, the page says an agent can write per-network versions, check limits, schedule and publish to real accounts, but every publish waits for your approval.

Which networks are supported?

X (Twitter), LinkedIn, Instagram, Facebook, Threads, Bluesky and YouTube Shorts, all from one place.

What is a credit?

One credit equals one published post per profile (five for X), one AI draft in the composer, or one analytics read of a post or a profile; an AI-generated image costs 20. Drafting inside your own AI over MCP costs nothing. The free plan includes 20 credits every month with no card required, and the free plan never expires.

Are social accounts safe?

Channels are authorized through each platform's official OAuth 2.0 sign-in, passwords are not seen or stored, tokens are encrypted at rest, and channels can be revoked instantly from the dashboard.

Can I connect my own AI or build on the API?

Both are possible. One command connects ChatGPT, Claude, Cursor or any MCP-capable client, and a REST API covers all seven networks. Neither is required to use the app.

Do I have to host or install anything?

For the app and for local AI clients, no: npx fetches and runs the MCP server on demand. Hosting is only needed for browser-based clients like Claude.ai and ChatGPT, which cannot spawn a local process. The server is MIT licensed and can be hosted anywhere that runs Node.

Can I cancel anytime?

Yes. The page states there are no contracts and no lock-in; cancellation is done from the billing tab.