You can now run SocialBooster straight from Claude and ChatGPT. We built a Model Context Protocol server, which is a fancy way of saying your AI assistant can talk to SocialBooster directly. Ask it to find a service, place an order, or check where an order is, and it does the work for you in the chat window.
No dashboard, no forms, no copy-pasting service IDs. You type what you want in plain language, the assistant handles the rest, and it confirms the cost with you before it spends a penny.
Here is what it does, why we built it, and how to connect it in about a minute.
What is an MCP server, in plain terms
The Model Context Protocol is an open standard for connecting AI assistants to real tools and data. Instead of an assistant guessing or making things up, MCP gives it a defined set of actions it can actually run, with real results coming back.
Our MCP server exposes a handful of SocialBooster actions to any assistant that supports the protocol. When you ask Claude "what is my balance?", it does not invent a number. It calls the real check_balance action, reads your actual prepaid balance, and tells you. Same for placing an order or checking status. The numbers are real because they come from the same system that powers the website and our REST API.
If you have used the SocialBooster API before, think of the MCP server as the same engine with a conversational front door.
What your assistant can do
The server ships with five tools. That is the whole surface, kept deliberately small and honest:
- list_services browses every service with live pricing and quantity limits, and you can filter it by platform.
- check_balance reads your prepaid SocialBooster balance in USD.
- place_order places an order from a service, a public link, and a quantity. It is paid from your balance.
- get_order_status tracks any order by its id and shows the delivery status.
- list_orders lists your most recent orders and their statuses.
Every one of these wraps the same API that runs the website. That means orders placed through Claude or ChatGPT inherit the same pricing, the same delivery, and the same order tracking as anything you place by hand. There is no separate, second-class path. It is the real thing, just driven by conversation.
What it looks like in practice
Say you want views on a TikTok. You would type something like:
Get me 5,000 TikTok views on tiktok.com/@me/video/123
The assistant checks the service list and your balance, then comes back with something like: "TikTok Views are $0.09 per 1,000, so 5,000 views is about $0.45. Your balance is $12.40, plenty to cover it. Shall I place the order?"
You say yes, it places the order, and it reports back: "Order placed and pending. Views start landing within minutes, and you can ask me to check the status anytime."
A few more things you can just ask for:
- "What SocialBooster services do you have for TikTok?"
- "Order 1,000 Instagram followers for instagram.com/mypage"
- "How much would 10,000 YouTube views cost?"
- "Check the status of my last order"
- "Show me my five most recent orders"
The point is that you never learn our menu structure or memorise a service id. You describe the outcome you want, and the assistant maps it to the right service and price before it acts.
Is it safe to let an AI place orders?
This is the first question everyone asks, and it is the right one. Here is the honest answer, with the guardrails laid out plainly.
You stay in control. Assistants that support MCP ask you to confirm before they run an action that spends money. You see the service, the quantity, and the cost, and nothing gets ordered until you say go.
It cannot overspend. Orders are paid from your prepaid SocialBooster balance. If the balance is not there, the order does not happen. There is no card on file for an AI to run away with, and no credit line to blow through.
It never sees your password. You authenticate with your API key as a Bearer token, exactly like the REST API. The server has no access to your login. The only thing you ever share with it about a post is the public link to the profile or content you want to grow, which is information anyone can already see.
It costs nothing extra. The MCP server is free to use. You pay only for the orders you place, at the same prices as the website and the API. There is no subscription and no per-message fee from us.
None of this removes your judgement from the loop. Read what the assistant proposes before you confirm it, the same way you would check an order form before hitting submit. The confirm step is there so a stray sentence never turns into a stray charge.
Which assistants work with it
Any client that supports MCP or custom connectors. In practice that covers:
- Claude on claude.ai and Claude Desktop
- ChatGPT through custom connectors
- Other MCP-compatible tools and clients
You connect it once with the server URL and your API key, and it stays available in that assistant until you remove it.
How to connect it in a minute
Step one: grab your API key. Create an API key in your SocialBooster dashboard. It is the same key the REST API uses. Keep it secret, because it spends from your balance. If you do not have an account yet, you can sign up free first.
Step two: add the connector. The exact path depends on your assistant:
- In ChatGPT, go to Settings, then Connectors, add a custom connector, paste the server URL, and set the Bearer token to your API key.
- In Claude on the web, go to Settings, then Connectors, and add a custom connector with the same URL and API key.
- In Claude Desktop, add the server to your config file using the
mcp-remotehelper, with your key passed as anAuthorization: Bearerheader.
The server URL and a ready-to-paste Claude Desktop config are both on the MCP page.
Step three: start growing. Ask the assistant to list services, then place an order. It confirms the details and the cost with you before anything is charged.
Why we built this
Two reasons, both honest.
The first is that this is genuinely where a lot of work is heading. People are running more of their day through an assistant, and asking that assistant to go and do small, well-defined tasks. Placing an SMM order is exactly that kind of task: pick a service, paste a link, set a quantity, confirm. It suits a conversational interface well, and forcing you back to a dashboard for it started to feel like friction we did not need to keep.
The second is that we already had the plumbing. Our REST API has done this job reliably for a long time, and the MCP server sits on top of it rather than beside it. That is deliberate. It means the AI path is not a risky new system with its own quirks. It is the proven one, reached a different way, with a confirmation step in front of anything that costs money.
If you have been automating with our API already, the MCP server is a lighter-weight option for the moments you do not want to write code. And if you have never touched the API, it is the easiest way yet to place an order without leaving the tool you are already talking to.
Connect it, ask it what it can do, and let it place your first order. The MCP page has everything you need to get started.