Eliza V2 Beta dropped today, March 17, 2025, and Phala Cloud is the first TEE-enabled platform to support its agent swarm deployment. For developers working with AI agents, this means you can now spin up ElizaOS-powered workflows with built-in security for sensitive operations—think API keys, on-chain transactions, and private communications—all handled in a TEE environment.

What’s the Deal with TEEs for AI Agents?

TEEs bring a layer of isolation that’s tough to beat. Unlike standard cloud setups, they keep your agent’s secrets (keys, credentials, etc.) locked down, even from the host environment. For Eliza V2, this unlocks some practical use cases:

• Twitter Delegation: Let a marketing agent post or interact without risking account creds.

• API Integrations: Call external services securely, no leaks.

• On-Chain Actions: Execute blockchain transactions via an investor agent with confidence.

• Private Messaging: Keep DMs or sensitive data encrypted end-to-end.

Phala Cloud integrates this TEE capability directly into the ElizaOS ecosystem, so you’re not just running agents—you’re running them with a security edge.

Deploy your first AI agent for free on Phala Cloud here.

How To Get Started

Eliza V2 agents can be deployed via the ElizaOS CLI or Phala Cloud CLI, both of which play nice with Phala’s TEE setup.

Here’s the breakdown:

1. CLI Deployment

If you’re comfortable with a terminal, the Phala Cloud CLI has you covered. It’s straightforward to set up and launch your swarm:

📖 Quickstart Guide

2. No-Code Option

Not into coding? Phala Cloud also offers a no-code path to deploy agents. It’s a solid choice if you want to experiment without digging into scripts:

📖 Deploy AI Agents Without Coding

3. Container Management View

Head to “Containers” section on Phala Cloud to view deployed containers.

4. Ready-to-deploy AI agent Swarms

Once the swarm is initiated, your agents will start to take directions from you and begin your journey building out your agent workflow.

Why This Matters for Developers

Most AI hosting platforms focus on ease of use, but security often takes a backseat. Phala Cloud flips that by leaning on TEEs to protect what matters—your agent’s keys, transactions, and data. It’s not about being the flashiest option; it’s about giving you a way to build and deploy agent swarms without sweating the usual risks.

An added advantage is that deploying multiple agent swarms within a single instance optimizes costs, making per-agent deployment significantly more efficient.

Check out the tweet from hashwarlock for a quick take:

About ElizaOS v2 and Phala Network

ElizaOS v2, as highlighted in the above tweet by Agent Joshua (@hashwarlock) on March 17, 2025, represents an exciting evolution of the autonomous agent platform designed to empower decentralized AI systems. ElizaOS v2 introduces a suite of enhanced features, including the ability to launch runtime environments, manage agent swarms, and onboard agents seamlessly. These advancements signify a leap forward in creating secure, multi-agent systems capable of operating autonomously across distributed networks. ElizaOS v2 is perfect for developers and organizations wanting to create smart, scalable apps that handle complex tasks without needing a central authority, positioning it as a key player in the fast-changing world of Web 3 technologies.

Phala Network, a leader in TEE-based computing solutions, complements ElizaOS v2 perfectly by providing the secure TEE infrastructure needed for such advanced agent systems. Phala, known for its decentralized cloud computing platform, updated its phala-cloud-cli on March 8, 2025, enhancing integration with its TEE-based cloud. This ensures computations run in isolated, tamper-proof environments, safeguarding data privacy and code integrity—critical for securely operating ElizaOS v2’s agent swarms. This synergy allows ElizaOS v2 to leverage Phala’s secure, trustless computing resources, enabling developers to deploy and manage autonomous agents with confidence, driving innovation in fields like blockchain, AI, and decentralized applications.