# evalview-mcp

Capture an AI agent’s current behavior and compare later runs against it. Useful for spotting changes to tool calls or answer quality after an update.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 140
- Author: hidai25
- License: Apache-2.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/hidai25/eval-view
- Updated: 2026-10-16
- HTML: https://affpapa.org/en/nevibecode/mcp/evalview-mcp

## What it does

EvalView records an agent run as a baseline, then checks later runs against it. The comparison covers tool names, parameters, and call order, so a prompt edit that changes the agent’s actions can show up even when the final response looks acceptable.

Tool and sequence comparisons run offline without an API key. You can add an LLM judge to score output quality, while running the agent itself may still incur provider charges. Setting up the server in Claude Code uses `OPENAI_API_KEY`, so have that key ready.

## Requirements
- Needs: Python and uv (commands run via uvx)
- Environment variables: `OPENAI_API_KEY` — keys and tokens: ask the human for the values, never commit them or put them in shared files.

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add evalview-mcp -e OPENAI_API_KEY=<OPENAI_API_KEY> -- uvx evalview
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from evalview <command>

### Codex

**Platform CLI**
```bash
codex mcp add evalview-mcp --env OPENAI_API_KEY=<OPENAI_API_KEY> -- uvx evalview
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from evalview <command>

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "evalview-mcp": {
            "command": "uvx",
            "args": [
                "evalview"
            ],
            "env": {
                "OPENAI_API_KEY": "<OPENAI_API_KEY>"
            }
        }
    }
}
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from evalview <command>

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "evalview-mcp": {
            "type": "stdio",
            "command": "uvx",
            "args": [
                "evalview"
            ],
            "env": {
                "OPENAI_API_KEY": "<OPENAI_API_KEY>"
            }
        }
    }
}
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from evalview <command>

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add -e OPENAI_API_KEY=<OPENAI_API_KEY> evalview-mcp uvx evalview
```
Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from evalview <command>

## When to use it
- compare agent runs with a saved baseline
- spot changes in tool-call order
- catch lower answer quality after a prompt edit
- check agent behavior before merging changes

## Example prompts
- Use `evalview-mcp` to compare my agent’s latest run with its saved baseline.
- Check whether the agent changed its tools or the order of its calls.
- Assess whether the agent’s answers lost quality after my last change.

## Similar
- [Front-End Checklist](https://affpapa.org/en/nevibecode/mcp/front-end-checklist.md) — MCP server · ★ 74559
- [Keploy](https://affpapa.org/en/nevibecode/mcp/keploy.md) — MCP server · ★ 18727
- [SeleniumBase MCP](https://affpapa.org/en/nevibecode/mcp/seleniumbase-mcp.md) — MCP server · ★ 13169
- [mobile-mcp](https://affpapa.org/en/nevibecode/mcp/mobile-mcp.md) — MCP server · ★ 9946
- [e2e](https://affpapa.org/en/nevibecode/mcp/e2e.md) — MCP server · ★ 11912
- [mockserver](https://affpapa.org/en/nevibecode/mcp/mockserver.md) — MCP server · ★ 4996

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md