Skip to main content
Glama
README.md
# revibe-mcp

MCP server for [Revibe](https://revibe.codes) codebase analysis. Works with Claude Desktop, Cursor, Windsurf, and any MCP-compatible client.

Analyze any GitHub repo — get architecture, file roles, execution flows, system design Q&A, and structured agent context.

## Quick Start (Node.js — zero install)

Add to your MCP client config:

```json
{
  "mcpServers": {
    "revibe": {
      "command": "npx",
      "args": ["revibe-mcp"]
    }
  }
}
```

Then ask your AI assistant to "log in to Revibe" — it will open your browser for a one-time signup.

## Quick Start (Python)

```bash
pip install revibe-mcp
revibe-mcp-auth login
```

```json
{
  "mcpServers": {
    "revibe": {
      "command": "revibe-mcp"
    }
  }
}
```

## Where to add the config

| Client | Config location |
|--------|----------------|
| **Claude Desktop** | `~/Library/Application Support/Claude/claude_desktop_config.json` |
| **Claude Code** | `.mcp.json` in your project root |
| **Cursor** | Settings > MCP Servers > Add |
| **Windsurf** | MCP config in settings |

## Auth

Two options:

**Option A: Browser login (recommended)**
```bash
# Node.js
npx revibe-mcp-auth login

# Python
revibe-mcp-auth login
```
Opens your browser — sign up or log in, click "Authorize". API key saved automatically to `~/.config/revibe/credentials.json`.

**Option B: Manual API key**

Get a key from [app.revibe.codes/settings](https://app.revibe.codes/settings), then pass it via env:
```json
{
  "mcpServers": {
    "revibe": {
      "command": "npx",
      "args": ["revibe-mcp"],
      "env": {
        "REVIBE_API_KEY": "rk_live_your_key_here"
      }
    }
  }
}
```

## Tools

| Tool | Description |
|------|-------------|
| `revibe_login` | Log in via browser — one-time setup |
| `analyze_repo` | Submit a GitHub repo for analysis (auto-detects from git remote) |
| `check_status` | Check progress of an analysis job |
| `get_summary` | Architecture, modules, patterns overview |
| `get_section` | Drill into architecture, file_roles, system_design_qa, etc. |
| `get_agent_context` | Full structured JSON optimized for AI agents |

## Usage

Once configured, just ask your AI assistant:

- "Analyze this repo with Revibe"
- "What's the architecture of github.com/user/repo?"
- "Get the agent context for this codebase"

The tools handle auto-detection, polling, and structured output automatically.

## CLI Commands

```bash
# Node.js
npx revibe-mcp-auth login     # Log in via browser
npx revibe-mcp-auth status    # Show current auth status
npx revibe-mcp-auth logout    # Remove saved credentials

# Python
revibe-mcp-auth login
revibe-mcp-auth status
revibe-mcp-auth logout
```

## Claude Code Skill

If you use [Claude Code](https://claude.ai/code), you can install Revibe as a slash command skill instead of (or in addition to) the MCP server:

```bash
# Copy the skill file
mkdir -p ~/.claude/skills/revibe
curl -o ~/.claude/skills/revibe/SKILL.md \
  https://raw.githubusercontent.com/selvatuple/revibe-mcp/main/skills/claude-code/SKILL.md
```

Then use `/revibe` or `/revibe github.com/user/repo` inside Claude Code.

The skill file is also available at [`skills/claude-code/SKILL.md`](skills/claude-code/SKILL.md).

## Packages & Distribution

| Method | Install |
|--------|---------|
| **npm** (zero install) | `npx revibe-mcp` |
| **PyPI** | `pip install revibe-mcp` |
| **MCP Registry** | [registry.modelcontextprotocol.io](https://registry.modelcontextprotocol.io) — search "revibe" |
| **Claude Code Skill** | Copy [`skills/claude-code/SKILL.md`](skills/claude-code/SKILL.md) to `~/.claude/skills/revibe/` |

## License

MIT

TDQS

A4/5.0

Scored across 6 tools

Disambiguation5/5

Each tool has a distinct role: login for authentication, analyze_repo for submission, check_status for job monitoring, and three retrieval tools with clear separation of scope. get_summary, get_section, and get_agent_context serve different granularities of output, preventing confusion.

Naming Consistency4/5

Most tools follow a clear verb_noun pattern (analyze_repo, check_status, get_summary, get_section, get_agent_context). The exception is revibe_login, which places a brand prefix before a verb and does not match the conventional verb-first structure, creating a minor inconsistency.

Tool Count5/5

With 6 tools, the server is well-scoped for its purpose: authentication, analysis submission, status polling, and result retrieval. This is an appropriate size that covers the workflow without redundancy or unnecessary bloat.

Completeness5/5

The tool set covers the complete lifecycle of a Revibe analysis: login, submit repository, check progress, and retrieve results in multiple formats (summary, sections, and full agent context). No critical steps are missing for the intended use case, and the output options are sufficient for both human and agent consumption.

Maintenance

ActivityInactive
ResponsivenessNo issues