Multiple projects are rolling out MCP protocols, gaining significant traction as the missing link to connect AI agents with the real world—and with each other.

What is MCP?
Still confused about MCP? Let’s break it down.
MCP (Multi-Agent Communication Protocol) is a standard introduced by Anthropic and now widely adopted by AI giants like OpenAI, Google, Cursor, and WindSurf. Think of it as a universal language for AI agents to communicate and collaborate.

1. DeMCP

Launched on April 25, DeMCP saw nearly eight-figure trading volume within its first hour. This real-time Web3 MCP marketplace offers discounted access to LLMs like GPT-4 and Claude. Built with Trusted Execution Environment (TEE) and on-chain security mechanisms, it has partnered with Phala Network for technical infrastructure.

2. Cookie.fun

Cookie.fun has debuted its dedicated MCP server:

• Plug-and-play MCP servers tailored for AI agents

• Designed for developers and non-technical users

• Zero configuration required

3. OpenServ

OpenServ’s MCP system is set to launch soon, enabling AI agents to integrate with real-world applications. Demo videos showcase compatibility with platforms like AWS, Discord, Notion, Figma, Google Maps, and Slack. As agents gain the ability to connect with everything, the possibilities become endless.

4. Holoworld

Holoworld is building multiple apps on the Open MCP platform. One standout example is Agent Studio, powered by AVA AI. The Holoworld app store is expected to expand further, with new MCP-driven applications in the pipeline.

5. arc

In talks with the arc team this week, we learned about their upcoming AI agent app store, Ryzone, which will deeply integrate with MCP. Notably, AI agents will leverage MCP servers from Rig Complex and Stripe to handle payments. Stripe, which processed $1.4 trillion in payments last year, adds serious credibility to this vision.

6. Heurist

Heurist now allows users to create custom MCP servers. This democratizes access to agent communication frameworks, empowering developers to tailor solutions for niche use cases.

7. ElevenLabs

ElevenLabs demonstrated how to build AI agents with personalized voices to order pizza via phone calls. Their MCP server connects to telephony systems and integrates Claude AI. Additionally, their WhatsApp message-control platform lets agents access chat histories and provide updates on demand. ElevenLabs also rolled out voice-to-text features for WhatsApp voice message transcription.

But Concerns Linger...
Ironically, an agent named Jail Breaker—famous for "jailbreaking" games—caught Elon Musk’s attention. Critics worry: Could MCP servers hide malicious code to trick agents into leaking data?

8. Freysa

Freysa proposes using Phala Network and NEAR Protocol to build encrypted wallet MCP servers. By securely integrating AI tools, this approach could minimize private key exposure risks.

The Future of MCP
MCP isn’t fading—it’s cementing itself as the gold standard for agent communication. As these eight projects push boundaries, the AI agent space may be on the brink of its next hype cycle. Will MCP-fueled agents redefine automation, or will security flaws derail the momentum? One thing’s clear: The race to connect AI to the real world just got a lot more interesting.