Best sandbox for LangChain and LangGraph agents in 2027

Q: What should a LangGraph agent's sandbox do?

Follow the thread. A sandbox named after the LangGraph thread_id lets the next turn find the same files, installed packages and running processes. Branch with the graph. When a graph explores two approaches, a copy of the running sandbox, including its memory, lets each branch start from the same state without rerunning its setup. Run what the model writes, not just Python. An in-process interpreter stops at Python. A microVM runs npm test, cargo build or a database. Keep the host out of it. A tool that calls subprocess runs on the machine serving your graph. A sandbox tool sends the command to a machine with its own kernel.

Q: When might another sandbox fit better?

LangSmith's own sandboxes. A team running everything inside LangSmith may prefer the sandboxes listed beside its tracing and deployments.

Which sandbox is best for LangChain and LangGraph?

It depends on where your graph already lives. For a shell and files that follow each thread, Runtime has the lowest cost of the sandboxes priced below: $22.78 for ten thousand runs, compared with $70.76 on Vercel Sandbox and $110.40 on Daytona. It bills measured CPU. A paused thread keeps its memory, and the same machine is a Deep Agents backend. Pick Bedrock AgentCore or Azure dynamic sessions if code must stay inside AWS or Azure. Pick LangSmith's sandboxes if your team runs everything in LangSmith.

Which sandboxes work with LangChain?

LangChain's docs, read on 25 October 2026, list these sandboxes:

Kind	Examples	What the model gets
Code interpreter tools	Amazon Bedrock AgentCore, Azure Container Apps dynamic sessions, Capsule	Runs a snippet and returns its output
Deep Agents sandbox backends	LangSmith, Daytona, E2B, Modal, Runloop, Vercel, AgentCore, NVIDIA OpenShell	A shell and files behind the harness

Any Python function can also be a tool. A sandbox's own SDK can therefore be wrapped as tools, with no extra layer in between.

What should a LangGraph agent's sandbox do?

Follow the thread. A sandbox named after the LangGraph thread_id lets the next turn find the same files, installed packages and running processes.
Branch with the graph. When a graph explores two approaches, a copy of the running sandbox, including its memory, lets each branch start from the same state without rerunning its setup.
Run what the model writes, not just Python. An in-process interpreter stops at Python. A microVM runs npm test, cargo build or a database.
Keep the host out of it. A tool that calls subprocess runs on the machine serving your graph. A sandbox tool sends the command to a machine with its own kernel.

What do LangGraph runs cost on each sandbox?

Ten thousand runs a month, each four minutes on a 2 vCPU, 4 GiB sandbox and busy for 40 CPU-seconds, at each provider's published rates:

Provider	Isolation	A month of runs	Runtime costs less by
Runtime Cloud	Firecracker microVM	$22.78
Vercel Sandbox	Firecracker microVM	$70.76	68%
E2B	Firecracker microVM	$110.40	79%
Daytona	Containers by default	$110.40	79%
Modal	gVisor, a shared kernel	$158.64	86%
Runloop	A microVM, container inside	$213.03	89%

When might another sandbox fit better?

LangSmith's own sandboxes. A team running everything inside LangSmith may prefer the sandboxes listed beside its tracing and deployments.

Runtime Cloud

Runtime's sandbox_tools(sbx) gives a LangChain agent or a LangGraph ToolNode a shell and a filesystem in a Firecracker microVM, through four plain functions. It is also a Deep Agents backend. RuntimeSandbox(sbx) from withruntime.deepagents implements BaseSandbox, so execute, grep and edit_file all act on the microVM (Deep Agents in a sandbox).

Name the sandbox after the thread_id with Sandbox.getOrCreate. It resumes from pause on the first tool call of the next turn. A paused sandbox costs no compute.
sbx.fork() copies the running sandbox, including its memory, for each branch.

The ten thousand runs above cost $22.78 on Runtime, 68% less than Vercel Sandbox and 79% less than Daytona, at rates checked 25 October 2026.

Runs on Runtime, recorded 1 October 2026

The same four functions serve both frameworks:

Pythonfrom langchain.agents import create_agentfrom langchain_core.tools import toolfrom withruntime import Sandboxfrom withruntime.tools import sandbox_toolswith Sandbox.create() as sbx:    agent = create_agent("claude-sonnet-4-6", tools=[tool(f) for f in sandbox_tools(sbx)])    result = agent.invoke({"messages": [{"role": "user", "content": "Fit a line to (1,2), (2,4.1), (3,6.2) and give the slope."}]})    print(result["messages"][-1].content)

We ran create_agent from LangChain 1.4.3, and a LangGraph 1.2.12 graph with a ToolNode and tools_condition. Each used a scripted chat model in place of Claude. The tool calls were fixed, and the frameworks ran them against live sandboxes:

textLangChain create_agent  runtime_write_file slope.py     Wrote 79 bytes to /workspace/slope.py  runtime_exec python3 slope.py   exit_code 0, stdout "2.1"LangGraph ToolNode  runtime_write_file fizzbuzz.py  Wrote 85 bytes to /workspace/fizzbuzz.py  runtime_exec python3 fizzbuzz.py    exit_code 0, stdout "1 2 Fizz 4 Buzz Fizz 7 8 Fizz Buzz 11 Fizz 13 14 FizzBuzz"

The slope came from NumPy, which is already in the image. The graph wiring and the per-thread sandbox are in LangGraph in a sandbox. The agent's options are in LangChain in a sandbox.

Run your graph on free usage after npx withruntime login. Tools that call E2B, Daytona or Vercel Sandbox today move with the migration guide.

Sources

Checked 25 October 2026.

LangChain tool integrations: the code interpreter integrations
Deep Agents sandboxes: the sandbox backend providers
Each provider's published rates, checked 25 October 2026, as the pricing guide lists them
The recorded runs: Runtime's Python SDK with LangChain 1.4.3 and LangGraph 1.2.12 on Python 3.12, against production on 1 October 2026