Skip to main content
Glama
Sign In
VegaMCP

VegaMCP

by Pastarafian
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

VegaMCP is a production-grade MCP (Model Context Protocol) server providing an autonomous AI agent swarm, persistent semantic memory, browser automation, multi-model reasoning, security gateway, agent graphs, zero-trust identity, A2A protocol, Docker sandbox v5.0, AI-first testing suite (mobile, web, API, desktop, accessibility, security, visual), and 78+ tools ā€” all accessible via any MCP-compatible client.

Version 7.2 (Sovereign Intelligence) introduces the production Claw Command Center, ultra-fast SQLite chat syncing, local vector Semantic Memory, LLM Output Evaluation, and 6 Unified Omni-Clusters.

šŸ“– Complete Features

Read FEATURES.md for a comprehensive list of all 17 unified V7 capability clusters including Docker Sandbox v5.0.

Related MCP server: Gravitas-Core-MCP

Quick Start

Prerequisites

  • Node.js 20+

  • npm 9+

Installation

# Clone the repository
git clone https://github.com/Pastarafian/VegaMCP.git
cd VegaMCP

# Install dependencies
npm install

# Copy environment config
cp .env.example .env
# Edit .env with your API keys

# Build
npm run build

Connect to VS Code (Gemini / Copilot)

Create .vscode/mcp.json in your workspace:

{
  "servers": {
    "REDACTED": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/VegaMCP/build/index.js"],
      "cwd": "/path/to/VegaMCP"
    }
  }
}

Note: API keys can be set in the env block of mcp.json or in the .env file (dotenv is loaded automatically).

Configuration

Copy .env.example to .env and configure:

# At least one reasoning model key required
OPENROUTER_API_KEY=          # Supports ALL models via OpenRouter
DEEPSEEK_API_KEY=            # Direct DeepSeek API (R1 + Chat)
KIMI_API_KEY=                # Kimi K2.5 for coding

# Optional integrations
GITHUB_TOKEN=                # GitHub API (60ā†’5000 req/hr)
TAVILY_API_KEY=              # AI-powered web search
SEARXNG_URL=                 # Self-hosted search fallback
SENTRY_AUTH_TOKEN=           # Error tracking
SENTRY_ORG=
SENTRY_PROJECT=

# Budget controls
TOKEN_DAILY_BUDGET_USD=5.00
TOKEN_HOURLY_BUDGET_USD=1.00

# Tool profiles
VEGAMCP_TOOL_PROFILE=full    # full | minimal | research | coding | ops

Project Structure

VegaMCP/
ā”œā”€ā”€ src/
ā”‚   ā”œā”€ā”€ index.ts                     # Server entry point + hub router
ā”‚   ā”œā”€ā”€ mcp-extensions.ts            # Sampling, logging, progress, roots
ā”‚   ā”œā”€ā”€ mcp-protocol/               # v6.0 / v7.0 protocol modules
ā”‚   ā”œā”€ā”€ db/                          # SQLite + vector store
ā”‚   ā”œā”€ā”€ swarm/                       # Agent swarm (10 agents)
ā”‚   ā”œā”€ā”€ tools/                       # All tool implementations
ā”‚   ā”œā”€ā”€ resources/                   # MCP resource providers
ā”‚   ā”œā”€ā”€ prompts/                     # MCP prompt templates
ā”‚   ā””ā”€ā”€ security/                    # Rate limiter, validator, guard
ā”œā”€ā”€ .env.example                     # Environment template
ā”œā”€ā”€ package.json
ā””ā”€ā”€ tsconfig.json

License

MIT

Available Tools

6 tools
omni_assistantC

Core AI intelligence: reasoning, budget, auto-update, prompts, web search (speed/balanced/quality modes + domain filter), code analysis, docs, sequential thinking, expert toolkits, knowledge engine, memory graph, semantic memory (vector search), context7 library docs.

Inputs: tool_target (required), payload (required).

omni_automationC

Action & control: browser, workflow, filesystem, shell, git, MCP tasks, API requests, sandbox, watchers, webhooks, schedule, notify, the_claw, claw_command_center, vps_control, image_generation.

Inputs: tool_target (required), payload (required).

omni_researchD

Research & Science: graph RAG, agentic RAG, self evolution, hypothesis gen, quality gate, synthesis engine, LLM router, memory bridge, tool discovery, sentinel, LLM output evaluation.

Inputs: tool_target (required), payload (required).

omni_swarmC

Agent orchestration: swarm ops, agent intel/ops, A2A protocol, agent graphs, agentic sampling, multimodal embeddings, session manager, MCP relay.

Inputs: tool_target (required), payload (required).

omni_systemsD

Infrastructure: database, postgres_client, health check, analytics, skills, seed data, sentry, OAuth, gateway, tool search, MCP apps, vault, dynamic indexing.

Inputs: tool_target (required), payload (required).

omni_testingC

QA & Security: mobile, web, API, DB, server, desktop, accessibility, security, visual, advanced testing, security scanner, stress test, zero trust.

Inputs: tool_target (required), payload (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelogā€¦

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    C2.2/5.0

    Scored across 6 tools

    Disambiguation1/5

    Tools have extremely broad and overlapping descriptions (e.g., omni_assistant and omni_research both cover RAG and knowledge; omni_automation and omni_swarm both involve orchestration). An agent would struggle to distinguish which tool to use for a given task.

    Naming Consistency4/5

    All tools follow a consistent 'omni_<noun>' pattern. However, the nouns (assistant, automation, research, swarm, systems, testing) are vague and do not clearly indicate the tool's unique scope.

    Tool Count2/5

    With only 6 tools covering an enormous range of capabilities (from AI assistants to infrastructure to testing), the set is under-scoped. Each tool effectively contains multiple distinct functionalities that should probably be separate tools.

    Completeness3/5

    The tool set covers many domains (AI, automation, research, orchestration, infrastructure, testing), but there are noticeable gaps such as dedicated monitoring, deep analytics, or security scanning beyond basic mentions. The overlap also indicates incomplete separation of concerns.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • Tyra Advanced Memory MCP

      bigrock776
      -
      license
      Not graded
      quality
      Not graded
      maintenance
      A sophisticated MCP server providing advanced memory capabilities with RAG, hallucination detection, and enterprise-grade AI infrastructure for intelligent agent ecosystems.
      -

    • Gravitas-Core-MCP

      ahmed-coding
      A
      license
      B
      quality
      D
      maintenance
      Production-grade, autonomous Model Context Protocol (MCP) server that elevates AI models from stateless code generators into persistent, self-verifying software engineers.
      21
      1
      MIT