Skip to content
LLM Cutover logo

LLM Cutover MCP server

Which MCP servers are LLM Cutover's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-11-03

← Back to LLM Cutover

Whose server it is
  • Registry namecom.llmcutover/auditverified in the MCP registry

The registry name sits under llmcutover.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-11-03
  • Stateanswerschecked 2026-11-03
  • Tools2
  • Server calls itselfllmcutover
  • Server version1.0.0
  • Protocol version2025-06-18

On 2026-11-03 the server answered initialize and tools/list without a sign-in. First answer in 59 ms.

Should an agent connect itNot proven either way yet
  • Its tools do every use case we checked: check whether a cheaper model matches your current model's answers; find a confidence threshold for routing calls to a cheaper model; estimate the monthly bill after routing calls to a cheaper model.
  • Checked on 1 of the 7 days it takes to judge reliability.

Once connected

  • For "check whether a cheaper model matches your current model's answers", the likely first call is run_cutover_audit.
  • For "find a confidence threshold for routing calls to a cheaper model", the likely first call is run_cutover_audit.
  • For "estimate the monthly bill after routing calls to a cheaper model", the likely first call is run_cutover_audit.
  • Parameters are untyped or undescribed in 1 of 2 tools: check arguments before calling.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 1, the first on 2026-11-03
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 1 so far
  • What an agent can do with it3 of 3 use cases
  • Answeredmeasuring1 of the 7 daily checks it takes to print a share
  • Median first answer116 ms
  • Space in an agent's context≈ 722 tokensthe definitions of all its tools (2), which an agent carries on every turn
  • Tool definitionsA · 92 of 100
  • Typed results0 of 2 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 1 of 2

What an agent can do with it

  • can do · check whether a cheaper model matches your current model's answers
  • can do · find a confidence threshold for routing calls to a cheaper model
  • can do · estimate the monthly bill after routing calls to a cheaper model

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers2 tools · read 2026-11-03
  • run_cutover_audit

    Measures how often a cheaper model (Sage 2) agrees with your current LLM on your samples, the confidence threshold for routing, and the monthly cost now vs a hybrid. Returns the full report as JSON.

  • list_models

    Current models with $/1M token prices and effort levels, usable as current.model.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "audit": {
      "type": "http",
      "url": "https://api.llmcutover.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http llmcutover https://api.llmcutover.com/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://api.llmcutover.com/mcp as a custom connector in the settings.

OpenAPI spec

4 paths, read 2026-11-03. Open the spec ↗

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.