Direct answer

Add an MCP Server to VS Code

Open the MCP configuration, add one remote server, and start it. In .vscode/mcp.json or the user MCP configuration, add a zambo entry under the servers key with type http and the url https://zambo.dev/mcp, save, start the server, and open the Chat view in agent mode: the Zambo tools are listed and callable. That one connection gives VS Code 153 tools in total, 35 on the default surface, memory shared with your other AIs, and a verifiable receipt for every tool call.

Updated October 20, 2026.

Why add an MCP server to VS Code at all

VS Code with Copilot in agent mode can already read your files, edit across a project, and run terminal commands. What it cannot do out of the box is reach past the workspace: pull a live price, audit a public web page, search for demand signals, save a fact so another AI can recall it next week, or file a record of a step that someone outside your machine can check later. An MCP server is how you close that gap. MCP, the Model Context Protocol, is the common wiring that lets an editor like VS Code offer outside tools to the model in the session, with the tools listed, named, and callable the same way built in tools are.

People search for how to add an MCP server to VS Code for two different reasons, and both land here. Some have a specific server in mind, often Zambo, and just need the exact file and fields. Others have a half working setup from another editor and are surprised that the block they pasted does nothing. VS Code has its own shape for this: its config file is called mcp.json, and the top level key inside it is servers, not the mcpServers key that Cursor, Cline, and Windsurf use. That one word is responsible for a large share of silent failures, so this page puts the snippet first and the explanation around it.

The config snippet

Create or open .vscode/mcp.json inside your workspace, or open the user level MCP configuration if you want the server in every workspace. Add this entry, keeping any servers you already have:

{
  "servers": {
    "zambo": {
      "type": "http",
      "url": "https://zambo.dev/mcp"
    }
  }
}

Three details carry the whole setup. The key is servers. The type is http, because Zambo is a hosted remote server, not a program VS Code launches on your machine. And the url is exactly https://zambo.dev/mcp, with no extra path and no API key, because the endpoint needs no key to connect. If you copied a block from another editor guide, expect to rename the top level key and drop any command, args, or env fields it carried: those belong to locally run servers, which Zambo is not.

Step by step, from a clean VS Code window

First, decide the scope before you touch a file. A workspace entry in .vscode/mcp.json travels with the project, can be committed for a team, and only applies when that folder is open. A user entry applies to every workspace on that machine and stays out of project repos. For a first test, the workspace file is easier to find and easier to delete, so start there. Second, open the Command Palette and run the MCP command to open or create the MCP configuration for the workspace, or create the .vscode folder and mcp.json file by hand; both routes land on the same file. Third, paste the snippet above and save. VS Code reads MCP configuration as JSON, so one missing comma anywhere in the file stops every server in it, not just the new one.

Fourth, start the server. VS Code shows a start control above the server entry in mcp.json, and MCP servers can also be started from the Command Palette and from the tools area of the Chat view. Click Start for zambo and watch for an error message rather than assuming silence means success. Fifth, open the Chat view in agent mode and open the tools picker: zambo should be listed, and expanding it should show its tools. A server that starts but shows no tools has usually connected to the wrong address, so recheck the url character by character. Sixth, run the first call described below before you fold Zambo into real work, because the receipt check at the end proves the whole chain, config, connection, account, tools, and record, in one small loop.

Verify the connection properly

A green state in the tools list is necessary but not sufficient. The real verification is a finished round trip: ask Copilot in agent mode to use Zambo for something small and checkable, for example a current price lookup or a quick audit of a public page you know well. The result comes back with a receipt link. Open the receipt page, copy its ID, and paste it into zambo.dev/verify. A match confirms the saved result was not changed after the fact, and it confirms something more practical too: your editor, the remote endpoint, and the public record all agree with each other. If the call works but no receipt appears, you are not looking at a Zambo tool result, so check that the agent picked the zambo tool and not a built in one with a similar name.

For a wider look at the catalogue, the same config entry with https://zambo.dev/mcp?full=true exposes all 153 tools instead of the 35 on the default surface. Treat full as a mode for audits, exploration, and wide agent runs, not as the daily default: in an editor session the model is also holding your codebase, and a lean tool list leaves it more attention for the code. Most VS Code users should run the plain endpoint day to day and switch the entry only for a session that genuinely needs the long tail.

Good first jobs inside VS Code

The fastest way to learn what the connection is for is to give the agent jobs that mix the workspace with the outside world. Ask it to audit the landing page your project serves and save the findings to memory before it edits anything. Ask it to check a live figure, a price, a version, a public claim, instead of answering from training memory, and keep the receipt with the change it justified. Ask it to package a finished piece of client work as a handover, so the delivery is a link with steps and receipts rather than a summary in a chat window. Each of those jobs uses a different part of the connection, tools, memory, receipts, handover, and each leaves something behind that a plain editor session would not.

What the connection unlocks

One connection changes four things at once, and it helps to keep them separate, because people usually come for the first and stay for the other three. Tools: your AI stops being limited to conversation and can run real jobs, lookups, checks, audits, file and data work, lead finding, and more, through the 153 tools on the Zambo endpoint, with 35 loaded on the default surface so the session stays focused. Memory: facts you explicitly save with memory_remember are held on your Zambo account and come back with memory_recall, in this client and in any other AI connected to the same account, so a decision saved today is present next week without you retyping the backstory. Continuity: work state saved in one AI can be opened in another, so hitting a limit or changing tools mid job means continuing, not restarting. Receipts: each saved tool result is filed with a verifiable receipt in the AER-1 format, draft-zambo-aer1-24, and anyone can check it free at zambo.dev/verify with no signup.

The flagship on top of those four is handover: when a job is done, the goal, the steps in order, the outputs, and the receipts travel in one link that a client or teammate opens with no account and accepts or flags. Try the shape before you rely on it: run a job end to end at the live demo, open the record it produces, then follow a recorded day at zambo.dev/today. The endpoint itself, https://zambo.dev/mcp, is the same one URL for every client on this page and every sibling guide, which is the point: connect once per client, keep one account, and let the work, the memory, and the proof accumulate in one place instead of scattering across apps.

What it costs

You can start on the free tier, and checking a receipt is always free. For $0.99, a Day Pass covers 24 hours of full use, framed for this page as one full day of connected work where every job you run leaves a record you can show someone. Ongoing use is the Zambo Pass at $29 a month or $290 a year, and $19 a month for $ZAMBO holders; see pricing for the current tiers. Humans pay by PayPal or card through PayPal; agents pay USDC on Base through x402. Nothing about the connection itself costs anything: the config entry, the command, and the first verification call are all free to try.

Troubleshooting

• Nothing happens after pasting the config. Check the top level key first: VS Code wants servers in mcp.json. A block pasted with mcpServers from another editor guide is valid JSON and silently wrong for VS Code.
• The server entry shows an error on start. Validate the whole file as JSON, confirm the type is http, and confirm the url is exactly https://zambo.dev/mcp with https, not http, and no trailing extra path.
• Zambo starts but lists no tools. Reload the VS Code window once, reopen the Chat tools picker in agent mode, and confirm you are looking at the workspace or user config you actually edited; editing one scope and testing in the other is the classic trap.
• A teammate does not see the server. A user level entry lives on your machine only. Commit the workspace .vscode/mcp.json if the team should share it, and expect each teammate to start it once in their own editor.
• Calls fail partway through a big session. The connection is fine if earlier calls returned receipts; a usage limit is the likelier cause, and a Day Pass or Zambo Pass raises it.

Honest limits

Frequently asked questions

Related

• Add an MCP server to Windsurf, the serverUrl version
• Add an MCP server to Codex, one command or config.toml
• Give your AI tools: what connecting an MCP server actually gives it
• Add Zambo MCP to Cline, step by step
• Add Zambo MCP to Zed, step by step
• Best MCP servers for VS Code, chosen by job