LLM Gateway

v1.0.0Coding Agentsstable

Comprehensive MCP server exposing dozens of capabilities to AI agents: multi-provider LLM delegation, browser automation, document processing, vector ops, and cognitive memory systems

agentllmmcpmcp-server
Share:
150
Stars
0
Downloads
0
Weekly
0/5

What is LLM Gateway?

LLM Gateway is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to comprehensive mcp server exposing dozens of capabilities to ai agents: multi-provider llm delegation, browser automation, document processing, vector ops, and cognitive memory systems

Comprehensive MCP server exposing dozens of capabilities to AI agents: multi-provider LLM delegation, browser automation, document processing, vector ops, and cognitive memory systems

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Comprehensive MCP server exposing dozens of capabilities to

Use Cases

Delegate requests across multiple LLM providers
Automate browser interactions and document processing
Manage vector operations and cognitive memory
LicenseNOASSERTION
Languagepython
Versionv1.0.0
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx llm-gateway-mcp-server

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use LLM Gateway

Ultimate MCP Server (LLM Gateway) is a comprehensive Python-based MCP server that exposes 40+ tools to AI agents, covering multi-provider LLM delegation (OpenAI, Anthropic, Google, DeepSeek, Grok, OpenRouter), browser automation via Playwright, document processing and OCR, SQL and vector database operations, a four-tier cognitive memory system, filesystem access, CLI tool wrapping, and audio transcription — all through a unified MCP interface with intelligent cost-optimized routing and semantic caching. It is designed to function as a complete AI agent operating system, potentially reducing API costs by 70-90% through intelligent task delegation and caching. Teams building complex multi-step AI agents use it as a single integration point that replaces dozens of individual MCP servers.

Prerequisites

  • Python 3.13 or later and the uv package manager (https://docs.astral.sh/uv/)
  • API keys for at least one LLM provider: OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, DEEPSEEK_API_KEY, OPENROUTER_API_KEY, or GROK_API_KEY
  • Playwright browser binaries for browser automation tools (installed separately)
  • Tesseract OCR engine if you plan to use PDF OCR capabilities
  • An MCP-compatible client such as Claude Desktop, Cursor, or a custom agent
1

Install uv and clone the repository

uv is the required package manager. Install it then clone the repository.

curl -LsSf https://astral.sh/uv/install.sh | sh
git clone https://github.com/Dicklesworthstone/ultimate_mcp_server.git
cd ultimate_mcp_server
2

Create a virtual environment and install dependencies

Use uv to create a Python 3.13 venv and install all optional extras.

uv venv --python 3.13
source .venv/bin/activate
uv sync --all-extras
3

Create a .env file with your API keys

Configure at minimum one LLM provider key and the server's host/port. All provider keys are optional — only configure the ones you need.

# .env
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=AIza...
DEEPSEEK_API_KEY=...
GATEWAY_SERVER_PORT=8013
GATEWAY_SERVER_HOST=127.0.0.1
GATEWAY_CACHE_ENABLED=true
GATEWAY_CACHE_TTL=86400
LOG_LEVEL=INFO
4

Start the MCP server

Launch the server using the umcp CLI. Use --include-tools or --exclude-tools to limit the active tool set for your use case.

umcp run
# Or with custom settings:
umcp run -h 0.0.0.0 -p 9000 -w 4
# Or exclude browser tools if Playwright is not installed:
umcp run --exclude-tools browser_init browser_navigate
5

Add the server to your MCP client configuration

Configure Claude Desktop or your agent to connect to the running Ultimate MCP Server via HTTP.

{
  "mcpServers": {
    "ultimate-mcp": {
      "type": "http",
      "url": "http://127.0.0.1:8013"
    }
  }
}
6

(Optional) Install browser and OCR dependencies

Enable Playwright browser automation and Tesseract OCR support with these additional installation steps.

# Playwright browser
python -m playwright install chromium

# OCR support (requires Tesseract engine on host)
uv pip install -e ".[ocr]"
# macOS: brew install tesseract
# Ubuntu: apt-get install tesseract-ocr

LLM Gateway Examples

Client configuration

Claude Desktop configuration connecting to Ultimate MCP Server running locally on the default port 8013.

{
  "mcpServers": {
    "ultimate-mcp": {
      "type": "http",
      "url": "http://127.0.0.1:8013"
    }
  }
}

Prompts to try

Representative agent tasks that exercise the server's multi-provider LLM delegation, memory, browser automation, and document processing tools.

- "Use the cheapest available model to summarise this 50-page PDF and store the result in episodic memory"
- "Navigate to https://news.ycombinator.com, extract the top 10 story titles, and save them to a file"
- "Run this SQL query against the connected database and return the results as a CSV"
- "Benchmark gpt-4o vs claude-3-5-sonnet on this classification task using 5 iterations each"
- "Search your semantic memory for anything related to our previous discussion about vector databases"
- "Extract all tables from this Word document and convert them to structured JSON"

Troubleshooting LLM Gateway

umcp command not found after installation

Ensure you have activated the virtual environment with: source .venv/bin/activate. The umcp CLI is installed inside the venv and is not available globally unless you install with uv tool install.

Browser tools fail with a Playwright not installed error

Run: python -m playwright install chromium inside the activated venv. Browser binaries are not installed by uv sync and must be downloaded separately. If you do not need browser tools, exclude them at startup: umcp run --exclude-tools browser_init browser_navigate

High API costs on first use

Enable semantic caching by setting GATEWAY_CACHE_ENABLED=true and GATEWAY_CACHE_TTL=86400 in your .env file. The cache deduplicates identical or semantically similar requests across sessions, which the project claims can reduce costs by 70-90% for repeated workloads.

Frequently Asked Questions about LLM Gateway

What is LLM Gateway?

LLM Gateway is a Model Context Protocol (MCP) server that comprehensive mcp server exposing dozens of capabilities to ai agents: multi-provider llm delegation, browser automation, document processing, vector ops, and cognitive memory systems It connects AI assistants to external tools and data sources through a standardized interface.

How do I install LLM Gateway?

Follow the installation instructions on the LLM Gateway GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with LLM Gateway?

LLM Gateway works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is LLM Gateway free to use?

Yes, LLM Gateway is open source and available under the NOASSERTION license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "llm-gateway-mcp-server": { "command": "npx", "args": ["-y", "llm-gateway-mcp-server"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use LLM Gateway?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides