Quorum with VS Code.

Wiring it up

Quorum is a hosted MCP server. There is no package to install and nothing runs on your machine.

.vscode/mcp.json  (workspace)  ·  or your user profile

{
  "inputs": [
    {
      "type": "promptString",
      "id": "quorum-key",
      "description": "Quorum API key",
      "password": true
    }
  ],
  "servers": {
    "quorum": {
      "type": "http",
      "url": "https://www.quorum.dog/mcp",
      "headers": { "Authorization": "Bearer ${input:quorum-key}" }
    }
  }
}

${input:...} makes VS Code prompt for the key on first use rather than storing it in a file you might commit — password: true masks it as you type and keeps it out of the settings file afterwards. You can hardcode the token instead, but a checked-in API key is a bad afternoon.

The inputs array is the half that is easy to leave out. ${input:quorum-key} is a reference; without a matching id declared alongside servers, there is nothing for VS Code to resolve it against and the placeholder is what gets sent as your bearer token. The server then answers 401 invalid_api_key, which reads as a bad key rather than an absent one and sends you hunting for a credential that was never wrong.

Get a key from your organisation dashboard.

What you get

All six are on the hosted server. The two testing tools spend real money on purpose — run_test bills per question at 25% off normal deliberate pricing and takes dry_run: true to price a run without spending anything; certify_mode commits $50 flat for 150 questions and runs for roughly two and a half hours, because the shared execution engine paces at about one question a minute. Do not call either speculatively.

estimate is worth more than its price row suggests. Alongside difficulty_score and task_type it returns deliberation_value — whether a panel is expected to help here, not just what it would cost. likely_helps on multi-step reasoning, likely_hurts when the answer is a passage to be reproduced rather than reasoned to (a synthesis pass rewrites by design, which corrupts it), unknown when there is no strong signal either way. It carries basis: "hypothesis" and its evidence line — LiveBench 2026-08-22, n=36, plus tier — so you can tell advice from measurement. The escalation router shows it in a branch.

The mistake to avoid

If you have copied an MCP config from anywhere else, check the top-level key first. VS Code expects servers. Claude Code and Cursor both use mcpServers, and a config carrying the wrong wrapper is not rejected with a useful message — the server simply never appears.

Everything inside the wrapper is the same across all three.

What it is good for here

VS Code's agent has your workspace; Quorum has whatever paragraph you hand it. That shapes what the pairing is actually for.

• Decisions with a migration behind them — schema changes, auth models, a public API shape. Being confidently wrong is expensive, and differently-trained models genuinely disagree here.
• Second-opinion on an answer you already have. Paste the proposed approach in and ask the panel to attack it. You want to know whether three models find the same weak spot, not to get a rewrite.
• Not for renames, formatting, or anything with one right answer. Your existing assistant is faster and better at those, and a panel adds a minute for nothing.

The paper

This page gets Quorum wired into VS Code. AI Deliberation & VS Code is the other half: when convening a panel here is worth the wait, the jobs it pays off on, and the ones it does not, with the exhibits behind each. 13 pages, PDF.

Every paper in the series sits on the whitepaper shelf.

Check it works

Ask your assistant to call list_modes. It is free, it requires a valid key, and it returns the panels you are actually entitled to — so one real answer proves the URL, the key and the wiring together.

curl -sS -X POST https://www.quorum.dog/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer qk_..." \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

A GET to that URL returns 405. That is correct, not a fault — the server is stateless, so there is nothing for a GET to stream.

Before you put it in a loop

Measured on production, 2026-08-20:

Asking what a question is costs effectively nothing. Convening the panel does not. The escalation router covers how to decide between them — and the decision is yours to make, not a threshold we set.

Other editors

Claude Code and Cursor nest servers under mcpServers. VS Code uses servers. Same protocol, same URL, different wrapper key — and the wrong one fails silently rather than telling you why.