Sandbox-Agent

Visit Tool

sandbox-agent is an AI Agents & Automation tool that runs coding agents in sandboxes. It allows control over HTTP, supporting Claude Code, Codex, OpenCode, Cursor, Amp, and Pi.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Yes

Skill level

Technical

About

What is sandbox-agent?

sandbox-agent is a server designed to run coding agents within isolated sandbox environments, controlling them remotely via HTTP. It provides a universal API to manage various coding agents like Claude Code, Codex, OpenCode, Cursor, Amp, and Pi, normalizing their proprietary interfaces into a single, consistent schema. This tool addresses challenges such as the need for isolated execution environments for AI, the diversity of agent APIs, and the ephemeral nature of agent sessions by streaming events for persistence. It can be run as an HTTP server or embedded with a TypeScript SDK, offering flexibility for developers to integrate it into their applications and deploy across various sandbox providers.

Best used for

Ideal for developers who need to run and control AI coding agents in secure, isolated environments, integrate diverse agent APIs, and persist session data. Especially valuable for those building applications that leverage AI for code generation and execution, requiring a robust and flexible remote control mechanism.

Common actions

run coding agents

control AI agents

manage agent sessions

integrate AI agents

stream agent events

open-sourcelow-code/no-codedeepfakecollaborationworkflowsautomated workflowgithub copilotface swapping"AI Agents"

Capabilities

Key features

Universal Agent API
Universal Session Schema
Runs Inside Any Sandbox
Server or SDK Mode
OpenAPI Spec
OpenCode SDK & UI Support
Lightweight Rust binary

Target Audience

developerstartup founder

Integrations

e2bdaytonamodalcloudflare-containersagent-computerdockervercel

Pricing & Plans

Open Source

Free

FAQs

Which coding agents are supported by sandbox-agent?

sandbox-agent supports a range of coding agents including Claude Code, Codex, OpenCode, Cursor, Amp, and Pi. It normalizes their APIs, allowing you to swap between them without needing to rewrite your integration code.

Can sandbox-agent be run locally or does it require a sandbox provider?

sandbox-agent offers flexibility, allowing you to run it locally for development purposes. For production environments, it can be deployed to various sandbox providers such as E2B, Daytona, Vercel Sandboxes, or your own infrastructure.

How does sandbox-agent handle session data persistence?

This SDK does not directly persist session data. Instead, it streams events in a universal JSON schema. This design allows you to integrate with external storage solutions like Postgres, ClickHouse, or Rivet to persist and replay session data as needed.

Why is sandbox-agent built with Rust?

Rust was chosen for sandbox-agent due to its ability to produce a single static binary, ensuring fast startup times and predictable memory usage. This makes it highly suitable for running within sandboxes or CI environments without the overhead of a large runtime.

Trending

Subcategories trending in AI Agents & Automation

AI Frameworks & Infra Chatbots & Conversational AI Workflow Agents Personal Assistants RAG & Document AI Voice Agents

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce