Skip to content
MCP server

TuringCorp

Second opinion for agents: send two options, get the better with a calibrated confidence

Install
Pricing
Paid
Platforms
API
Runs
Remotely
Version
1.0.3
Listing updated
1 day ago

About TuringCorp

TuringCorp is a paid MCP server for ai models and llm tooling. It runs as a hosted remote server, so there is nothing to install locally.

TuringCorp exposes one tool, decide, that takes a neutral task and two options and returns which is preferred, a confidence percentage and a reason. The README says a panel of models judges the pair together. It is meant for calls with no objective answer, such as two plans, two drafts or two vendors, and it is explicit about when not to use it: more than two options, anything computable or verifiable, factual lookups, speed-critical paths and irreversible high-stakes calls without review.

The site publishes calibration figures: on JudgeBench accuracy runs from 99.6% in the 90%-and-up band down to 67.7% below 70%, and on ContextualJudgeBench from 83.3% down to 55.4%. Agents are advised to act on high values and escalate low ones.

A decision is a long call, so clients should allow 180 to 300 seconds. A job_id lets you collect a paid result for seven days instead of retrying. Tool listing needs no credentials but calling decide requires an Agent Pass sent as a bearer token.

Install

Run in a terminal

claude mcp add --transport http decider https://mcp.turingcorp.net/mcp

Add to Settings → Connectors → Add custom connector

https://mcp.turingcorp.net/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "decider": {
      "url": "https://mcp.turingcorp.net/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "decider": {
      "type": "http",
      "url": "https://mcp.turingcorp.net/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "decider": {
      "serverUrl": "https://mcp.turingcorp.net/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "decider": {
      "type": "http",
      "url": "https://mcp.turingcorp.net/mcp"
    }
  }
}

Tools this server provides (1)

Taken from the project documentation; the running server may differ.

decide
Pick the better of two options for a task, with confidence and a reason.

What it does

Pairwise judging
Compares exactly two options and names the preferred one.
Calibrated confidence
Returns how far apart the options were judged, with published accuracy bands.
Reasoned output
Each answer comes with a reason.
Retrievable results
A job_id collects a paid decision for seven days.
Stateless transport
Streamable HTTP with no session state.
Credential-free discovery
The tool list can be read without a pass.

Good for

  • Choose between two drafts of an email
  • Let an agent pick between two plans and escalate close calls
  • Compare two vendor proposals with a second judgement
  • Route on confidence instead of guessing

Pricing

Paid

Requires an Agent Pass; prices not published here

Prices change. Check mcp.turingcorp.net before you buy.

TuringCorp FAQ

How many options can it compare?

Exactly two. It will not rank a list.

How long does a call take?

Reserve 180 to 300 seconds; it is a long call set in the client or host.

What if the call times out?

Do not retry; retrieve the decision with the job_id for up to seven days.

How do I authenticate?

Calling decide needs an Agent Pass as a bearer token; listing tools does not.

Is TuringCorp free?

No. TuringCorp is a paid product.

How do I add TuringCorp to Claude Code?

Run this command in your terminal: claude mcp add --transport http decider https://mcp.turingcorp.net/mcp

What are the best alternatives to TuringCorp?

The closest alternatives to TuringCorp are Claude Code Ultimate Guide, signafAIde, Toolkien, Lugh and HypePaper.

Alternatives to TuringCorp

  1. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  2. signafAIde

    MCP server

    Local MCP server to set up, verify and check coverage of signed agent action records

    Runs locally
    • 5 tools
    • Freemium
    • Security
  3. Toolkien

    MCP server

    One MCP gateway for all your MCP servers, with code mode, permissions and audit

    Not checked yet
    • 3 tools
    • Remote
    • Open source
    • AI models and LLM tooling
  4. Lugh

    MCP server

    Write, check and dry-run Lugh agent programs with policy, budgets and approvals

    Runs locally
    • 5 tools
    • Open source
    • AI models and LLM tooling
  5. HypePaper

    MCP server

    Research papers, benchmarks, citations and topic experts for AI assistants

    • 58 tools
    • Remote
    • Freemium
    • AI models and LLM tooling