09/23/2025 · Shaunak Joshi
Introducing Remote MCP Support in Beta on GroqCloud
Connect to any tool. Share context seamlessly. All OpenAI compatible. Run faster at lower cost.
Today we’re announcing that remote Model Context Protocol (MCP) server integration is available in Beta on GroqCloud, unlocking faster, lower-cost AI applications with tool capabilities through Anthropic’s open MCP standard. Developers can now connect any remote MCP server of their choice to models hosted on GroqCloud, allowing models to interact with external tools (GitHub, browsers, databases, and more) via the OpenAI-compatible Groq Responses API.
Because our implementation is compatible with both the OpenAI Responses API and the OpenAI remote MCP specification, developers already running on OpenAI can switch to Groq with zero code changes and immediately benefit from Groq’s speed and predictable costs.
Why Remote MCP Matters
The Model Context Protocol is an open standard for connecting AI assistants to the systems where data lives (example: content repositories, business tools, and development environments). MCP defines how an AI agent can securely discover and invoke external tools, exchange context, and orchestrate multi-step actions.
In essence, MCP provides a universal interface to thousands of tools, transforming AI agents from isolated language models into powerful, connected systems.
GroqCloud now brings this capability to our low latency platform. The result:
- Faster agent execution: lower round-trip latency than alternatives.
- Lower operational costs: same experiences at a fraction of the price.
- Seamless migration: keep your connector code, just change the endpoint.
Supported Models at Launch
Remote MCP will be available on all GroqCloud models that support tool use (excluding Compound and Compound-Mini at launch):
- GPT-OSS-120B
- GPT-OSS-20B
- Llama 4 Maverick
- Llama 4 Scout
- Llama 3.3 70B
- Llama 3.1 8B
- Kimi K2 0905
- Qwen 3 32B
Key Features
| Feature | What It Means for Developers |
|---|---|
Drop-in compatibility | Keep your OpenAI Responses and MCP integration, just point it to GroqCloud. |
Speed | Groq’s deterministic inference means tool-using agents respond faster and more reliably. |
Lower operating cost | Run the same workloads more cost effectively at scale. |
Approved tool use | Clear allowlists and approvals to control which tools agents can call. |
Cookbook Tutorials to Help You Get Started
To help you get started with MCP support on GroqCloud, we have partnered with several tool providers to publish cookbook tutorials showcasing how to leverage their MCP servers with models hosted on GroqCloud. Including:
- BrowserBase MCP: Equip models via Groq API with Browserbase's MCP server for web automation using natural language commands, including tools for web interaction and data extraction.
- BrowserUse MCP: Equip models via Groq API with Browser Use's tools to enable autonomous website browsing, information extraction, and web pages interaction.
- Exa MCP: Use Exa's web search and web crawling tools to get real-time information from the internet. Find relevant search results, extract data from websites, and run deep-research.
- Firecrawl MCP: Equip GPT-OSS 120B via Groq API with enterprise-grade web scraping capabilities, intelligent data extraction, structured parsing, and deep web research.
- HuggingFace MCP with Groq: Retrieve real-time HuggingFace model data with the HuggingFace MCP and Groq API.
- Parallel MCP with Groq: Real-time search with access to live data with the Parallel MCP and Groq API.
- Stripe MCP: Automate your invoicing process with Stripe's MCP and Groq API.
- Tavily MCP with Groq: Build a real-time research agent with the Tavily MCP and Groq API.
Pricing & Packaging
You pay only for tokens consumed by the GroqCloud model you select. Bring your own MCP server and API key, any fees from that third-party server are billed directly by the provider.
Getting Started
Switch your endpoint to GroqCloud, choose a supported model, and configure your preferred MCP server, no code changes required. We can’t wait to see what you build when standard MCP meets Groq performance. Learn more about MCP support on GroqCould.