GitHub - sumant1122/opscat: 🐾 Zero-config, single-binary Software Catalog, Interactive 2D Dependency Graph & native MCP Engine for Platform Engineering

🌟 Why OpsCat?

Legacy developer portals (like Backstage) are notoriously heavy, requiring weeks of boilerplate TypeScript setup, complex yarn monorepos, and high memory footprints (>1GB RAM).

OpsCat reimagines platform engineering for 2026:

• ⚡ Single Binary (< 30MB RAM): Written in Go with embedded React/Vite Web UI. Zero external runtime dependencies.
• 🔍 Zero-Config Auto-Discovery: Crawls your codebase to automatically infer tech stacks (Go, Node, Python, Rust, Java), containers (Dockerfile), infrastructure (K8s, Helm, Terraform), and APIs (OpenAPI, GraphQL, Proto).
• 🧠 Semantic Natural Language Search: Query your platform architecture in plain English ("Find services handling Stripe payments") via CLI, Web UI, or MCP tools.
• 🕸️ Interactive 2D Dependency Graph: Visualizes software relationships, systems, and resource dependencies in real time.
• 🤖 Native MCP (Model Context Protocol) Server: Gives AI coding agents (Claude, Cursor, Antigravity, Devin) direct context to query service architecture, owners, and health metrics.
• 🛡️ Automated Compliance Scorecards: Automatically calculates service health scores (0-100%) based on documentation, ownership, containerization, and API specs.

🚀 Quickstart

1. Installation

Build directly from source:

git clone https://github.com/opscat/opscat.git
cd opscat
go build -o bin/opscat ./cmd/opscat

2. Scan a Workspace Directory

Scan your current workspace to auto-discover services and save to SQLite:

3. Launch the Web Portal & MCP Server

Run a single command to serve the Glassmorphic Web UI, REST API, and MCP Server:

./bin/opscat serve --http :8080

Open http://localhost:8080 in your browser!

💻 CLI Commands

OpsCat provides an intuitive command-line interface:

🤖 AI Agent Integration (MCP)

OpsCat implements the Model Context Protocol (MCP) specification, allowing AI agents to understand your platform topology.

Configure with Cursor / Claude Desktop

Add to your MCP server configuration (mcpServers section):

{
  "mcpServers": {
    "opscat": {
      "command": "/usr/local/bin/opscat",
      "args": ["serve", "--stdio"]
    }
  }
}

Available MCP Tools for AI Agents

• catalog_semantic_search: Perform natural language semantic search across platform architecture, tech stacks, and APIs.
• catalog_list_services: List microservices with filtering by owner, tech stack, or minimum score.
• catalog_get_service: Retrieve detailed metadata, scorecards, and infrastructure specs.
• catalog_scan_workspace: Dynamically trigger workspace discovery.
• catalog_get_dependencies: Query upstream and downstream service dependencies.

📄 Manifest Spec (catalog.yaml)

OpsCat works out of the box with zero configuration. Optionally add a catalog.yaml to any repository root to override or enrich metadata:

kind: service
name: payment-service
title: Global Payment Gateway Service
description: Core PCI-compliant payment processing engine.
owner: team-fintech
system: checkout-pipeline
domain: e-commerce
tags:
  - payments
  - fintech
  - pci-dss
links:
  - title: GitHub Repository
    url: https://github.com/org/payment-service
    type: github
  - title: Grafana Dashboard
    url: https://grafana.internal/d/payments
    type: dashboard
relations:
  - type: dependsOn
    target: auth-service
  - type: consumesApi
    target: stripe-gateway

🏗️ Architecture Overview

flowchart TD
    subgraph Consumers[" 💻 Clients & Interfaces "]
        UI["🌐 Glassmorphic Web UI<br/>(Embedded React + Vite)"]
        CLI["💻 Terminal CLI<br/>(opscat scan / query / serve)"]
        MCP["🤖 AI Coding Agents<br/>(Cursor / Claude / Antigravity via MCP)"]
    end

    subgraph Engine[" ⚡ OpsCat Core Engine (Single Go Binary < 30MB) "]
        subgraph Ingestion[" 🔍 Auto-Discovery Engine "]
            LangDetector["Languages & Frameworks<br/>(Go, Node, Python, Rust, Java)"]
            InfraDetector["Container & Infrastructure<br/>(Docker, K8s, Helm, Terraform)"]
            APIDetector["API Specifications<br/>(OpenAPI, GraphQL, gRPC)"]
        end

        GraphStore[("💾 Unified Entity Graph Store<br/>(In-Memory + SQLite)")]
        ScorecardEngine["🛡️ Compliance Scorecard Engine<br/>(Automated 0-100% Health Scores)"]
        
        RESTServer["📡 REST API Router<br/>(/api/v1/entities, /api/v1/graph)"]
        MCPServer["🔌 Native MCP Server<br/>(Stdio & HTTP SSE Transports)"]
    end

    Ingestion --> GraphStore
    GraphStore --> ScorecardEngine
    ScorecardEngine --> RESTServer
    ScorecardEngine --> MCPServer

    RESTServer --> UI
    RESTServer --> CLI
    MCPServer --> MCP

🧪 Testing

Run unit tests across all core packages:

📜 License

Distributed under the Apache 2.0 License. See LICENSE for more information.