Skip to main content
The ZeroEval MCP server lets AI agents inspect traces, manage judges and prompts, submit feedback, run optimizations, and deploy to production, all without leaving the agent context. It speaks the Model Context Protocol, so any MCP-compatible client (Cursor, Claude Code, Windsurf, etc.) can connect directly.

​
Setup

The fastest way to get started is to point your MCP client at the hosted server. No installation required.

​
Cursor

Add this to your Cursor MCP settings (.cursor/mcp.json): 
{
  "mcpServers": {
    "zeroeval": {
      "url": "https://mcp.zeroeval.com/mcp",
      "headers": {
        "Authorization": "Bearer <your-project-api-key>"
      }
    }
  }
}

​
Claude Code

claude mcp add zeroeval --transport http https://mcp.zeroeval.com/mcp \
  --header "Authorization: Bearer <your-project-api-key>"

​
Other MCP clients

Any client that supports HTTP transport works. Set the server URL to https://mcp.zeroeval.com/mcp and pass your project API key in the Authorization: Bearer <key> header.
Get your project API key from the ZeroEval dashboard under Settings → API Keys.

​
Resources

The server exposes two MCP resources for introspection:
URIDescription
config://server-contextRedacted server config: auth mode, base URL, project scope, and feature flags
docs://capabilitiesCanonical tool and resource inventory with annotations and output contract summary

​
Tools

​
Read tools

Read tools are safe to call at any time. They do not modify state.
ToolDescription
list-tracesList recent traces
get-traceGet a trace with its spans
list-judgesList all judges
get-judgeGet judge details and linkage state
list-judge-evaluationsList evaluations from a judge
get-judge-criteriaGet scoring criteria for a judge
list-promptsList all prompts
get-promptGet a prompt at a specific version or tag
list-prompt-versionsList all versions of a prompt
list-optimization-runsList optimization runs for a task
get-optimization-runGet run details with candidate prompt and metrics
get-project-summaryHigh-level project monitoring summary

​
Write tools

All write tools require confirm: true in the request and are annotated with destructiveHint: true so MCP clients can prompt for user approval before calling.
ToolDescription
create-judgeCreate a new judge
link-judge-to-promptLink a judge to a prompt
unlink-judge-from-promptRemove a judge’s prompt link
create-judge-feedbackSubmit feedback on a judge evaluation
create-prompt-feedbackSubmit feedback on a prompt completion
start-prompt-optimizationStart a prompt optimization run
start-judge-optimizationStart a judge optimization run
cancel-optimization-runCancel a running optimization

​
Deploy

Production deploys always require two steps:
  1. Preview: Call preview-optimization-deploy with the run ID. This verifies the run succeeded, summarizes the candidate vs current production, and returns a time-limited confirmation receipt.
  2. Deploy: Call deploy-optimization-run with confirm: true and the receipt from preview. The server re-reads current state and rejects the deploy if anything drifted since the preview.
ToolDescription
preview-optimization-deployPreview what deploying a run would do (read-only)
deploy-optimization-runDeploy a succeeded run to production (requires receipt + confirm)

​
Proposal tools

Proposal tools are read-only helpers that gather evidence or prepare the exact next mutating call without executing it.
ToolDescription
investigate-prompt-issuesGather evidence about prompt state and recommend next steps
investigate-judge-issuesGather evidence about judge state and recommend next steps
prepare-prompt-optimizationPropose the exact start-prompt-optimization call to make
prepare-judge-optimizationPropose the exact start-judge-optimization call to make

Tracing quickstart

Get your first trace in under 5 minutes

Prompt setup

Add ze.prompt() to your codebase

Judges

How calibrated judges evaluate your production traffic

CLI

Manage traces, prompts, and judges from your terminal