Direct answer

Add an MCP Server to ChatGPT

To add an MCP server to ChatGPT, open the MCP or connectors settings in ChatGPT, add a custom connector, and paste the server URL. For Zambo, that URL is https://zambo.dev/mcp, with no API key. Start a new chat, confirm the tools are available, and run one small call you can check. This page takes a complete beginner from that paste to working tools, memory, handover, and receipts.

Updated October 20, 2026.

What an MCP server is, in plain words

MCP stands for Model Context Protocol. In plain words, it is a standard way for an AI app to plug in outside tools, the same way a browser plugs in extensions or a phone installs apps. ChatGPT on its own can only talk. ChatGPT with an MCP server connected can also do: look things up live, run checks, work with files and data, save facts for later, and produce a record of what it did. The server is a separate service that sits at one web address. Your AI app connects to that address, reads the list of tools the server offers, and calls those tools when your request needs one.

That one idea explains the whole setup on this page. There is no software to install on your computer for a hosted server like Zambo, no code to write, and no command line to open. You tell ChatGPT the address of the server once, in its settings. From then on, every new conversation can use the tools at that address. If you have never added an MCP server before, this is a good first one to learn on: the address is short, there is no key to manage, and you can test the connection with one small call and check the result yourself in under a minute.

Before you start: what you need

You need three things, and nothing else. First, your ChatGPT account, signed in as usual. Second, the server address, which for Zambo is https://zambo.dev/mcp, copied exactly, with no extra words, spaces, or paths added. Third, about two minutes and one small test job you can verify by eye, such as checking a current price or auditing a page you know well. You do not need an API key. You do not need a developer account for this connection. You do not need to install anything, restart your computer, or change any other setting in ChatGPT.

One habit is worth building from the first minute: treat every new connection as unproven until you have run a test call and checked its result. A settings screen that saves without an error is encouraging, but it is not proof. The proof is a tool actually running, a result actually coming back, and, with Zambo, a receipt you can open and verify. This page builds that test into the steps, so by the end you will know the connection works, not just hope it does.

Step by step: add the Zambo MCP server to ChatGPT

That is the entire install. If you are used to software setups with downloads, permissions screens, and restarts, the shortness of this list can feel suspicious. It is short because the server is hosted: Zambo runs the service at that address, keeps it updated, and adds new tools on its side. Your side of the bargain is one accurate paste and one honest test. When Zambo ships a new tool, your next new chat simply has it, with nothing to update on your machine.

Your first tool call, and how to check it

In the new chat, ask for something small, current, and checkable. Ask ChatGPT to use Zambo to look up a current price you can compare against a source you trust, or to audit a web page you own and already know the state of. Watch for two things. First, the answer should come from a tool call, not from the model answering from memory: current data in the result is the giveaway. Second, the result should carry a receipt link, the record of that call.

Open the receipt. Then go to zambo.dev/verify and run the free check on it, with no signup. The check confirms the saved result was not changed after the fact. If the check matches, the full path works: ChatGPT reached the server at zambo.dev/mcp, the tool ran, the result was saved, and the record of it can be verified by anyone, including a client or a teammate who was not in the chat. Before connecting, you may want to see the same loop without any setup at all: the live demo runs one real Zambo call and hands you its receipt in about 10 seconds.

What you can do once ChatGPT is connected

Tools. The connected client can call Zambo tools directly in the session: web and data checks, current prices instead of training data guesses, site audits, document and file work, lead finding with the leadsignal tool, and the everyday jobs that used to mean opening five tabs and copying results back into the conversation by hand. Ask for the job in plain words and let the client pick the tool, then read the result the way you would read any work product, critically and in full.

Memory. Zambo memory holds only facts you explicitly save, through memory_remember, and brings them back through memory_recall. Save a preference, a project rule, a price, or the current state of a long job once, and any AI connected to the same Zambo account can recall it later. The store does not import your conversations and it does not guess: if a fact is in there, someone put it there on purpose, and you can review and correct it the same way.

Handover. When a job is finished, package it as a handover: one link carrying the goal, the steps in order, the outputs, and any open items. The person receiving it opens the link with no account and clicks Accept or Flag. You cannot accept your own handover, which is exactly what makes the acceptance recorded on it worth showing later.

Receipts. Every tool call returns a verifiable receipt for its saved result. Anyone can check a receipt free at zambo.dev/verify with no signup, and the check confirms the saved result was not changed after the fact. It does not prove the result was right, and this page will not pretend otherwise. It proves what ran and what came back, unaltered, which is the half of accountability that a confident summary cannot supply on its own.

Default surface or full surface: which URL should you paste?

The address in the steps above, https://zambo.dev/mcp, loads the default surface: 35 tools, chosen as the set most sessions actually reach for. A shorter tool list is a feature for everyday work. ChatGPT chooses tools faster, its choices are easier to follow, and you are less likely to watch it pick a near identical tool that was not the one the job needed. For research, drafting with live checks, file work, memory, and handover, the default surface is the right starting point and, for most people, the permanent one.

The full catalogue is one small change away: use https://zambo.dev/mcp?full=true and the session can see all 153 tools. Full makes sense on exploration days, when you do not yet know which specialist tool a job needs, and for deliberate audits across the whole set. It is rarely the best everyday setting. A practical pattern is to keep your normal connector on the default URL, and only switch to the full URL for the specific session that needs a tool outside the default set. Choice discipline here is the same as with any toolbox: the bench you work at daily should hold the tools you use daily.

Using the connection well in everyday ChatGPT work

A connection only earns its place if it changes what your sessions can finish. Three patterns tend to stick. First, live facts instead of stale ones: when an answer depends on a current price, a current page state, or current public information, ask for a tool call and read the result, rather than accepting a confident answer built from training data. Second, save as you go: when a session produces a decision, a format, or a fact you will need again, ask ChatGPT to save it with memory_remember in one sentence. Third, close jobs with a record: when work is for someone else, finish with a handover link and its receipts, so the other person checks the work instead of taking your summary on trust.

Notice what these patterns have in common: none of them require you to learn tool names or syntax. You ask in plain words, the client picks the tool, and you judge the output. The learning curve for an MCP connection is not in the using, it is in the trusting, and the receipt check is what makes that trust reasonable. Run the check a few times early on, until verifying feels as normal as reading the answer itself.

Troubleshooting: when the server does not show up

• The tools are not listed in a new chat. Reopen the MCP or connectors settings in ChatGPT and confirm the custom connector is saved, with the server URL exactly https://zambo.dev/mcp. One wrong character, an added space, or an extra path is the usual cause.
• The connector saved but calls fail. Start a completely new chat and try the small test call again. Existing conversations can predate the connector, so a fresh chat is the reliable test.
• You pasted the full URL by accident. https://zambo.dev/mcp?full=true is valid, it just loads all 153 tools. If tool choices look noisy, edit the connector back to the plain default URL.
• A call hits a usage limit. Checking receipts stays free at https://zambo.dev/verify/. For a full day of work, a Day Pass covers 24 hours of full access; see the cost section below.
• You want to remove the server. Delete the custom connector in the same settings where you added it. Removal is instant and reversible: re-adding is the same one paste.

What it costs

Connecting the server costs nothing to set up, and checking a receipt is always free. When the connection becomes part of a real working day, a Day Pass at $0.99 covers 24 hours of full access: enough to connect in the morning, run real jobs through the day, save the facts that matter, and hand the finished work over with its record attached. Ongoing use is the Zambo Pass at $29 a month or $290 a year, and $19 a month for $ZAMBO holders. Humans pay by PayPal or card through PayPal; agents pay USDC on Base through x402. Read that pricing as the price of solving the connection problem properly, with tools, memory, and proof in one place, rather than as a charge for the paste itself, which takes a minute and costs nothing.

Honest limits

Frequently asked questions

Related

• Add memory to ChatGPT that works in every AI you use
• Best MCP servers for ChatGPT, chosen by job
• Add an MCP server to Claude, step by step
• Add an MCP server to Cursor, step by step
• Cross-AI memory: facts that follow you from one AI to the next
• MCP server with receipts for every tool call