templates8 production templates · boot 50–100 ms · BYO Dockerfile

Every agent stack, pre-baked.

Each template is a Firecracker snapshot built from a Dockerfile or captured from a running sandbox, ready to restore in 50–100 ms. CPU and memory are baked into the template — pick one that fits or build your own. The vCPU count below is what the guest sees; sustained CPU depends on your plan, with Free capped at 2 cores per sandbox and Pro and Team bursting to all 8.

boot measured on the host; your network round trip is extra

8 of 8 shown

--template base

Base

Ubuntu 24.04 + mise with Node 24, Python 3.12, Go, and Bun pre-installed. The universal runtime that powers Apps.

runtime
node 22 · python 3.12 · go · bun
base
ubuntu:24.04 + mise
cpu
8 vCPU burst
memory
4 GB
size
2.0 GB
boot
50–100 ms

# included

node 22python 3.12gobunpnpmyarn
$ pandastack sandbox create --template base
docs →

--template code-interpreter

popular

Code Interpreter

Sandboxed Python + Node for data analysis, plotting, agent loops, and Functions. The default for LLM code execution.

runtime
python 3.11 · node 22
base
python:3.11 + node 22
cpu
8 vCPU burst
memory
2 GB
size
12 GB
boot
50–100 ms

# included

pandasnumpyscipyjupyterplaywrightopenai-agents
$ pandastack sandbox create --template code-interpreter
docs →

--template agent

new

Coding Agent

One box with every terminal coding agent baked in — claude, codex, opencode, amp, grok, gemini, copilot. Drop the matching API key and start a long-running agent.

runtime
claude · codex · opencode · amp · grok · gemini · copilot
base
ubuntu:24.04 + node 24
cpu
8 vCPU burst
memory
2 GB
size
6.0 GB
boot
50–100 ms

# included

claudecodexopencodeampgrokgeminicopilotripgrepfdgit
$ pandastack sandbox create --template agent
docs →

--template claude-agent

new

Claude Managed Agents

Run Anthropic's Claude Managed Agents on PandaStack. The ant worker executes the agent's tool calls inside an isolated microVM — Claude drives the loop, your sandbox runs the code. Sub-second starts, stateful multi-turn.

runtime
ant · node 22 · python 3.12
base
ubuntu:24.04 + ant + mise
cpu
8 vCPU burst
memory
2 GB
size
8.0 GB
boot
50–100 ms

# included

antnode 22python 3.12gitripgrep
$ pandastack sandbox create --template claude-agent
docs →

--template browser

Browser

Headless Chromium + Playwright + Xvfb, plus the Python crawl4ai extraction stack. Scrape, automate, screenshot, and extract.

runtime
chromium · playwright · crawl4ai
base
ubuntu:24.04 + chromium
cpu
8 vCPU burst
memory
4 GB
size
4.0 GB
boot
50–100 ms

# included

chromiumplaywrightcrawl4aitrafilaturaxvfbffmpeg
$ pandastack sandbox create --template browser
docs →

--template postgres-16

PostgreSQL 16

Managed Postgres 16 with pgvector and pgbouncer on a durable volume. Backs the Databases feature.

runtime
postgresql 16 · pgvector
base
ubuntu:24.04 + PGDG
cpu
8 vCPU burst
memory
1 GB
size
12 GB
boot
50–100 ms

# included

postgresql 16pgvectorpgbouncer
$ pandastack sandbox create --template postgres-16
docs →

--template custom

byoi

Custom build

Bring a Dockerfile or build script. We bake the rootfs snapshot and serve it from your registry.

runtime
anything that runs on linux
base
your image
cpu
8 vCPU burst
memory
you decide
size
you decide
boot
depends on size

# included

FROM any imageRUN any commandsnapshot baked once
$ pandastack sandbox create --template custom
docs →

--template snapshot-template

new

Live Snapshot

Start from any running sandbox, install tools interactively, then save it with sandbox.snapshot({ name: "..." }).

runtime
whatever you installed
base
running sandbox
cpu
inherited
memory
inherited
size
captured state
boot
50–100 ms

# included

sandbox.snapshotinteractive bakeinstant restore
$ pandastack sandbox create --template snapshot-template
docs →

Bring a Dockerfile or snapshot a live sandbox.

Bake an image once into a Firecracker rootfs + snapshot, or install tools inside a running sandbox and save that exact state as a template.

custom build
Dockerfile
FROM node:22-bookworm
RUN apt-get update && apt-get install -y --no-install-recommends \
git ripgrep fd-find tmux && rm -rf /var/lib/apt/lists/*
RUN npm i -g pnpm@9 && \
npm i -g @anthropic-ai/claude-code
WORKDIR /workspace
ENV NODE_ENV=production
pandastack template build
# Build the snapshot (~30s one-time bake)
# CPU and memory are baked into the template — set them here.
$ pandastack template build \
--name my-claude-stack \
--file Dockerfile \
--memory-mb 4096
# Restore a sandbox from it in 50–100 ms — it boots with 4 GB RAM (8 vCPU burst).
$ pandastack sandbox create \
--template my-claude-stack \
--ttl 1h

01

One-time bake

Build from Dockerfile layers or capture exact live sandbox state.

02

Instant restore

Every create restores from the baked snapshot — no cold boots after the first bake.

03

Reproducible

Snapshots are content-addressed. Pin a sha256 for production.

Don't see your stack?

Open an issue or send a PR — most community templates land within 48h.