Live 58 Actions

E2B MCP Server
for AI Agents

Q: What AI frameworks and AI clients does the StackOne E2B MCP server support?

The StackOne E2B MCP server supports both. MCP clients (paste-and-go apps): Claude Desktop, Claude Code, Cursor, VS Code, Goose. Agent frameworks (code SDKs you build with): OpenAI Agents SDK, Anthropic, Vercel AI, Google ADK, CrewAI, Pydantic AI, LangChain, LangGraph, Azure AI Foundry.

Connect your AI agent to StackOne's E2B MCP server and give it 58 MCP tools out of the box. Auth, tool execution, and security all managed.

Start Free Book Demo Read Docs →

E2B MCP Server

Built by

StackOne

Coverage

58 Agent Actions

Create, read, update, and delete across E2B — and extend your agent's capabilities with custom actions.

MCP Actions → Build Actions →

Authentication

Agent Tool Authentication

Per-user OAuth in one call. Your E2B MCP server gets session-scoped tokens with zero credentials stored on your infra.

Agent Auth →

Security

Agent Protection

Every E2B tool response scanned for prompt injection in milliseconds — 88.7% accuracy, all running on CPU.

Prompt Injection Defense →

Performance

Max Agent Context. Min Cost.

Free up to 96% of your agent's context window to enhance reasoning and reduce cost, on every E2B call.

Tools Discovery →

What is the E2B MCP Server?

A E2B MCP server lets AI agents read and write E2B data through the Model Context Protocol — Anthropic's open standard for connecting LLMs to external tools. StackOne's E2B MCP server ships with 58 pre-built actions, fully extensible via the Connector Builder — plus managed authentication, prompt injection defense, observability, and agent execution runtime. Connect it from MCP clients like Claude Desktop, Claude Code, Cursor, Goose, and VS Code, or from agent frameworks like OpenAI Agents SDK, LangChain, and Vercel AI SDK.

All E2B MCP Tools

Every action from E2B's API, ready for your agent. Create, read, update, and delete — scoped to exactly what you need.

Files

Upload File
Upload a file into a running sandbox, creating parent directories as needed.
Download File
Download a file from inside a running sandbox by path.
Move File
Move or rename a file or directory inside a running sandbox.
Remove File
Delete a file or directory inside a running sandbox.

Watchers

Create Watcher
Create a pollable filesystem watcher on a directory inside a running sandbox.
Remove Watcher
Destroy a filesystem watcher and free its server-side event buffer.

Sandboxs

Create Sandbox
Start a new sandbox from a template and return its ID and envd access token.
Get Sandbox
Get a single sandbox by ID, including its state, resource allocation, expiry and envd access token.

Sandbox Metrics

Get Sandbox Metrics
Return the CPU, memory and disk time series for one sandbox.
List Sandbox Metrics
Return metrics for several sandboxes in a single call.

Snapshots

Create Snapshot
Capture a sandbox's current state as a persistent snapshot that outlives the sandbox.
List Snapshots
List the team's snapshots, optionally filtered by source sandbox or name.

Secrets

Create Secret
Create a secret by storing a runtime marker as its first version.
List Secrets
List the project's secrets, returning metadata only.
Get Secret
Get one secret's metadata by identifier or name.
Update Secret
Rotate a secret by appending a new version of its stored marker.
Delete Secret
Revoke a secret and schedule its versions for cleanup.

Templates

Create Template
Create a new sandbox template and its initial build record.
List Templates
List the team's template definitions.
Update Template
Update a template's settings, such as whether it is public.
Delete Template
Delete a template and its builds.

Volumes

Create Volume
Create a new persistent volume for the team.
List Volumes
List the team's persistent volumes.
Get Volume
Get one persistent volume by ID.
Delete Volume
Delete a persistent volume from the team.

Other (33)

Send Process Input
Send stdin or PTY bytes to a running process inside a sandbox.
Send Process Signal
Send a termination signal to a running process inside a sandbox.
List Directory
List the entries of a directory inside a running sandbox, optionally recursing to a given depth.
Get Watcher Events
Drain the filesystem events a watcher has buffered since the last poll.
List Processes
List the processes currently running inside a sandbox.
Get Sandbox Health
Probe whether a sandbox's in-sandbox agent is alive and reachable.
Get Sandbox Service Stats
Return internal service statistics from a sandbox's in-sandbox agent.
List Sandbox Environment Variables
Return the environment variables visible inside a running sandbox.
List Sandboxes
List the team's sandboxes, newest first, with optional state, template, metadata and start-time filters.
Get Sandbox Logs
Read a sandbox's log stream with cursor, direction, level and search filters.
List Template Tags
List all tags pointing at builds of one template.
Get Team Metrics
Return the team's concurrent-sandbox and sandbox-start-rate time series over an interval.
Get Team Metrics Max
Return the peak value of one team metric over an interval.
List Template Builds
List all builds for one template.
Get Build Upload Link
Get a presigned upload link for a tar file of template build layer files.
Get Build Status
Get the status of one template build, with a slice of its logs.
Get Build Logs
Read a template build's log stream with cursor, direction, level and source filters.
Update Process
Update a running process inside a sandbox, such as resizing its pseudo-terminal.
Set Sandbox Timeout
Overwrite a sandbox's time to live so it expires the given number of seconds from now.
Update Sandbox Network
Replace the egress network rules on a running sandbox.
Delete Tags
Delete one or more tags from a template.
Stat File
Return metadata for one path inside a running sandbox.
Make Directory
Create a directory inside a running sandbox.
Compose Files
Concatenate several files inside a sandbox into one using zero-copy composition.
Close Process Stdin
Close standard input on a running process to signal end of file.
Connect To Sandbox
Return sandbox details and resume the sandbox first if it is paused.
Pause Sandbox
Snapshot a sandbox in place and stop it so it can be resumed later.
Kill Sandbox
Terminate a sandbox immediately and release its resources.
Fork Sandbox
Checkpoint a running sandbox and create new sandboxes from that snapshot.
Refresh Sandbox
Extend a sandbox's time to live to keep it alive.
Assign Tags
Assign one or more tags to a template build.
Check Template Alias
Check whether a template exists under a given alias.
Start Build
Start building a template image from a base image or template and a list of steps.

Set Up Your E2B MCP Server in Minutes

One endpoint. Any framework. Your agent is talking to E2B in under 10 lines of code.

MCP Clients

Agent Frameworks

Claude Desktop-

JSON

Claude Desktop Docs

{
  "mcpServers": {
    "stackone": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://api.stackone.com/mcp?x-account-id=<account_id>",
        "--header",
        "Authorization: Basic <YOUR_BASE64_TOKEN>"
      ]
    }
  }
}

Check More AI & ML MCP Servers

Hugging Face

126+ actions

OpenAI

121+ actions

Dialpad

57+ actions

Exa

35+ actions

Google Gemini

30+ actions

Anthropic

29+ actions

Fireflies.ai

25+ actions

All AI & ML MCP Servers →

Platform Resources

MCP Code Mode: Keeping Tool Responses Out of Agent Context

Anthropic's code_execution processes data already in context. Custom MCP code mode keeps raw tool responses in a sandbox. 14K tokens vs 500.

11 min

Comparing BM25, TF-IDF, and Hybrid Search for MCP Tool Discovery

Benchmarking BM25, TF-IDF, and hybrid search for MCP tool discovery across 916 tools. The 80/20 TF-IDF/BM25 hybrid hits 21% Top-1 accuracy in under 1ms.

10 min

Indirect Prompt Injection Defense for MCP Tools: A Technical Guide

MCP tools that read emails, CRM records, and tickets are indirect prompt injection vectors. Here's how we built a two-tier defense that scans tool results in ~11ms.

12 min

MCP vs A2A: Architecture, Security, and When to Use Each

MCP vs A2A: what each protocol standardizes, how they differ, their shared security risks including indirect prompt injection, and when to use one, both, or a hybrid architecture.

12 min

MCP vs API: What Hundreds of Connector Builds Taught Us

MCP wraps APIs, it doesn't replace them. After building hundreds of connectors that serve both, here's when each approach wins.

14 min read

E2B MCP Server FAQ

Does StackOne have a E2B MCP server?

Yes. StackOne offers a hosted E2B MCP server with 58 pre-built actions, and every action is tested and QA'd by StackOne. Connect it to Claude, Cursor, and any other MCP client, or to any agent framework through the AI Action SDK. It ships with managed agent authentication, prompt injection defense, and tool discovery with server-side execution that preserve your agent's context window and keep reasoning performance.

E2B MCP server vs direct API integration — what's the difference?

A E2B MCP server and direct API integration serve different use cases. Direct API integration is for software-to-software — backend code calling E2B. A E2B MCP server is for AI agents — MCP clients like Claude and Cursor, plus framework agents built with OpenAI, LangChain, or Vercel AI — discovering and calling E2B at runtime. StackOne provides both.

How does E2B authentication work for AI agents?

E2B authentication for AI agents works through a StackOne Connect Session. Create one via the dashboard or the SDK — you get an auth link and ready-to-paste config for Claude Desktop, Cursor, and other MCP clients. Your user authenticates their own E2B account; StackOne handles token exchange, storage, and refresh. Credentials never reach the LLM, and each user is isolated via origin_owner_id.

Are E2B MCP tools vulnerable to prompt injection?

Yes — E2B MCP tools can be vulnerable to indirect prompt injection. Any tool that reads user-written content — documents, messages, tickets, records, or free-text fields — is a potential vector. StackOne Defender scans every tool response before it enters the agent's context — regex patterns in ~1ms, then a MiniLM classifier in ~4ms. 88.7% accuracy, CPU-only.

What is the context bloat of a E2B agent and how do I avoid it?

Context bloat happens when E2B tool schemas and API responses eat your E2B agent's memory, preventing it from reasoning effectively. A single E2B query can return a massive JSON response, and connecting multiple tools compounds the problem. Tools Discovery and Code Mode reduce context bloat — loading only relevant tools per query and keeping raw responses out of the agent's context.

Can I limit which actions my E2B agent can access?

Yes — you can limit which actions your E2B agent can access directly from the StackOne dashboard. Toggle actions on or off, or restrict them to specific accounts, with no code changes to your agent. Session tokens can be scoped to exact actions so if one leaks, exposure stays contained.

Can I create custom agent actions for my E2B MCP server?

Yes — you can create custom agent actions for your E2B MCP server using Connector Builder. It's an integration agent your coding assistant (Claude Code, Cursor, or Copilot) can invoke to research E2B's API, generate production-ready connector YAML, test against the live API, and validate before you ship.

When should I NOT use a E2B MCP server?

Skip a E2B MCP server if your integration is purely software-to-software — direct E2B API integration is simpler when no AI agent is involved. For deterministic, compliance-critical operations (financial transactions, regulatory reporting), direct API gives you predictable behavior without agent-driven decision-making. MCP shines when AI agents need to dynamically discover and call E2B actions at runtime.

What AI frameworks and AI clients does the StackOne E2B MCP server support?

The StackOne E2B MCP server supports both. MCP clients (paste-and-go apps): Claude Desktop, Claude Code, Cursor, VS Code, Goose. Agent frameworks (code SDKs you build with): OpenAI Agents SDK, Anthropic, Vercel AI, Google ADK, CrewAI, Pydantic AI, LangChain, LangGraph, Azure AI Foundry.

Put your AI agents to work

All the tools you need to build and scale AI agent integrations, with best-in-class connectivity, execution, and security.

Start Free Book Demo

E2B MCP Serverfor AI Agents