Akamai Agent Sandbox

Open-source Spin component that runs agent-generated code in WebAssembly isolation with limits on compute, memory and network.

Akamai · Security and safety tooling · open-source

Published in Feed

Overview

Akamai Agent Sandbox, announced October 27, 2026 and in beta, is an open-source component of the Spin framework that uses WebAssembly isolation to cap the compute, memory, outbound connections and system capabilities of code that AI agents generate. Agents run on Akamai Functions and the sandbox is reached through the Spin SDK. It launched alongside Akamai Inference, the managed model-serving service.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • WebAssembly isolation for agent-generated code
  • Configurable limits on compute, memory, outbound connections and system capabilities
  • Open-source component of the Spin framework, used through the Spin SDK
  • Agents execute on Akamai Functions

Use cases

  • Running untrusted LLM-written code without exposing the host or network
  • Giving coding agents an execution tool with enforced resource ceilings
  • Edge-deployed agent tool execution

Pricing

Open source; beta. Akamai's press release does not state hosting pricing.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
27 Oct 2026
Record updated
2 Nov 2026
Record last checked
2 Nov 2026

Explore the Feed

Other entries filed under security and safety tooling; inclusion does not imply equivalent features.

  • DefenseClaw Edge Connector

    Cisco's open-source DefenseClaw gains an Edge Connector that allows or blocks AI agent tool calls directly on IoT and robot hardware.

  • Lakera Guard 2.0

    Lakera's real-time guardrail API for LLM apps, updated with multi-agent protection and a new benchmark for detecting indirect prompt injection in tool outputs.

  • Fleming-1

    Sierra's model that listens to a live phone call and scores whether the caller is an AI agent rather than a human.

Browse the Feed