# Pubroot MCP server

> Enables AI agents to search AI-reviewed research papers, verify factual claims, and retrieve structured peer review metadata and contributor reputation.

- Repository: https://github.com/buildngrowsv/pubroot-website
- Page: https://hostedmcp.org/servers/buildngrowsv/pubroot-website
- Transport: remote
- Stars: 0
- Latest version: 0.2.0

## Install

### Claude Code

```sh
claude mcp add --transport http pubroot-website https://mcp.pubroot.com/mcp
```

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

```json
{
  "mcpServers": {
    "pubroot-website": {
      "url": "https://mcp.pubroot.com/mcp"
    }
  }
}
```

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

```json
{
  "servers": {
    "pubroot-website": {
      "type": "http",
      "url": "https://mcp.pubroot.com/mcp"
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"pubroot-website","type":"http","url":"https://mcp.pubroot.com/mcp"}'
```

### Codex

```sh
codex mcp add pubroot-website --url https://mcp.pubroot.com/mcp
```

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

```toml
[mcp_servers.pubroot-website]
url = "https://mcp.pubroot.com/mcp"
```

### Gemini CLI

```sh
gemini mcp add -s user --transport http pubroot-website https://mcp.pubroot.com/mcp
```

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

```json
{
  "mcpServers": {
    "pubroot-website": {
      "url": "https://mcp.pubroot.com/mcp",
      "type": "http"
    }
  }
}
```

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `GITHUB_REPO` | no | no |  |
| `REPO_MODE` | no | no |  |

## FAQ

### What can an AI agent do with the Pubroot MCP server?

Agents can search for papers by keyword or topic, verify specific factual claims, retrieve full structured reviews, and check contributor reputation scores.

### How are the papers reviewed?

Submissions undergo a 6-stage AI review pipeline that includes novelty checks, repository analysis, and claim verification using web search grounding.

### What is the acceptance threshold for a paper?

To be accepted and auto-published, a submission must achieve a score of 6.0 or higher out of 10.0.
