# EvalView MCP server

> Performs snapshot testing for AI agents by recording tool-calling trajectories and flagging behavioral regressions.

- Repository: https://github.com/hidai25/eval-view
- Page: https://hostedmcp.org/servers/hidai25/eval-view
- Transport: stdio
- Stars: 142
- License: Apache-2.0
- Latest version: 0.6.0

## Install

### Claude Code

```sh
claude mcp add --env 'OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>' eval-view -- uvx evalview
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "eval-view": {
      "command": "uvx",
      "args": [
        "evalview"
      ],
      "env": {
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "eval-view": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "evalview"
      ],
      "env": {
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"eval-view","type":"stdio","command":"uvx","args":["evalview"],"env":{"OPENAI_API_KEY":"<YOUR_OPENAI_API_KEY>"}}'
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "eval-view": {
      "command": "uvx",
      "args": [
        "evalview"
      ],
      "env": {
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

### Codex

```sh
codex mcp add eval-view --env 'OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>' -- uvx evalview
```

### Codex config.toml (`~/.codex/config.toml`)

```toml
[mcp_servers.eval-view]
command = "uvx"
args = ["evalview"]

[mcp_servers.eval-view.env]
OPENAI_API_KEY = "<YOUR_OPENAI_API_KEY>"
```

### Gemini CLI

```sh
gemini mcp add -s user -e 'OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>' eval-view uvx -- evalview
```

### Gemini settings.json (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "eval-view": {
      "command": "uvx",
      "args": [
        "evalview"
      ],
      "env": {
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `OPENAI_API_KEY` | no | yes | OpenAI API key for LLM-as-judge output quality scoring. Optional — deterministic tool/sequence evaluation works without it. |

## FAQ

### How does EvalView differ from assertion-based testing?

Instead of writing manual assertions or metrics, EvalView records your agent's current behavior and flags any drift from that baseline, catching unanticipated regressions.

### What does the tool actually compare during a check?

It diffs the entire trajectory, including tool names, parameters, and the order in which they were called, rather than just the final output string.

### Does running EvalView require an LLM API key?

The deterministic tool and sequence comparison runs offline without an API key, though using an LLM judge for output quality scoring or executing the agent itself may incur charges.
