Introduction

​

Overview

In a world where AI Agents orchestrate complex workflows, on-chain logic and code execution must be as seamless as a smart contract call.

AVM’s MCP Server Model Context Protocol bridges LLMs and on-demand compute, delivering isolated runtime environments on a global network of peer-operated nodes.

AVM transforms code execution from a centralized service into a permissionless, composable layer. Agents talk to MCP endpoints; MCP endpoints invoke AVM nodes; AVM nodes return verifiable outputs—all without intermediaries.

​

Core Advantages

• Decentralized Execution: Agents route tasks through a distributed mesh of AVM nodes, ensuring censorship-resistance and fault tolerance.
• MCP-Native Integration: Plug-and-play with Claude, Cursor, and any client supporting the Model Context Protocol.
• Trust-Minimized Operations: Each code run is sandboxed, auditable, and governed by smart-contract–backed policies.
• Enterprise-Grade Performance: Autoscaled, GPU-ready environments optimized for data science, cryptography, and machine learning workloads.

​

Get Started

1. Generate an API Key
   Visit the API Keys page and create a new key to authenticate Sandbox Runs API calls.
2. Configure Your Agent
   Point your MCP client to the AVM server:

   {
     "mcpServers": {
       "avm": {
         "command": "npx",
         "args": ["@avm-ai/avm-mcp"],
         "env": { "AVM_API_KEY": "YOUR_KEY_HERE" }
       }
     }
   }
   

3. Launch Jobs
   Send your first code task via the Runs API or through your LLM’s toolchain and watch AVM return structured outputs.