Early Access — free while in beta

Know what your
AI agents are doing.

Caged runs coding agents in isolated Firecracker VMs. Every file edit, terminal command, and LLM call is recorded, costed, and scored — automatically.

No credit card · 5 sandboxes/day free · Works with Claude Code, Cursor, Aider, any MCP agent

caged — dashboard

3 active

sandboxes

$1.42

today's spend

87

trust score

auth-refactor
Claude Code$0.6892
fix-tests
Aider$0.3188
api-docs
Cursor$0.4379

Built for trust, not just execution.

Other platforms give agents a computer. Caged gives you full visibility and control over what they do with it.

Real VMs, not containers.

Every sandbox gets its own Linux kernel via Firecracker microVMs. Hardware-enforced isolation — an agent can't escape to your host or other sandboxes.

Every action recorded.

File edits, terminal commands, LLM calls, network requests — all captured with sub-millisecond timestamps. Replay any session from start to finish.

Budget caps that bite.

Set a dollar limit. When the agent hits it, the sandbox dies — no overruns. Track real-time spend per model, per session, per day.

Trust scores, not vibes.

Quantified risk per session. Attempted sudo? Score drops. Accessed .env? Score drops. Ran tests? Score goes up. Transparent and explainable.

Workflow

From zero to running agent in under a minute.

1

Create a sandbox.

Pick a template, set CPU/RAM limits, define a budget. One API call, one CLI command, or one click.

terminal
$ caged run --template node-20 \
    --repo github.com/acme/app \
    --budget 5.00

⚡ Sandbox sbx_7f3k ready (247ms)
🔗 ssh [email protected]
2

Connect your agent.

Point Claude Code, Cursor, Aider, or any MCP-compatible agent at the sandbox. Full Linux environment with your code pre-loaded.

Claude Codecompatible
Cursorcompatible
Aidercompatible
OpenHandscompatible
Custom MCPcompatible
3

Observe and control.

Watch live or replay later. Costs tick up in real-time. Trust score drops? Kill it instantly. Budget exceeded? Auto-killed.

14:02:31git clone acme/app
14:02:33npm install
14:03:01Claude 4 Sonnet — 2.4k tokens
14:03:04wrote src/auth.ts (+42 lines)
14:03:06npm test — 12 passed
14:03:08attempted: cat .env

Capabilities

Everything an agent needs.
Nothing it shouldn't have.

VM-Level Isolation

Dedicated kernel per sandbox. Hardware-enforced boundaries — real VMs, not container namespaces.

Real-Time Costing

Track every token and compute second as it happens. Hard caps auto-kill sessions that overspend.

Session Replay

Scrub through file edits, terminal commands, and LLM calls. Structured playback of entire sessions.

Trust Scoring

Flags credential access, destructive commands, and network anomalies. Quantified risk per session.

Any Agent

Claude Code, Cursor, Aider, OpenHands, custom MCP servers. If it speaks SSH or MCP, it works.

Scoped API Keys

Read-only keys for dashboards, full-access for CI. Create, rotate, revoke — zero downtime.

Cost Analytics

Per-model breakdowns, daily trends, top spenders. Understand spend before the invoice arrives.

Network Policy

Allowlist domains your agent needs. Block everything else. Reach npm — not your production DB.

Policy Engine

Define rules per tool, file, command, or network call. SOC2 and HIPAA templates built in. Policies inherit org → team → project.

Human-in-the-Loop

Pause execution when an agent hits a guardrail. Approve or reject from dashboard, Slack, or email. SLA timers auto-escalate.

Agent Eval

Define test scenarios in YAML. Run agents against assertions. Detect flakiness and regressions across runs.

Sub-Second Boot

Sandboxes launch in <300ms via Firecracker. No cold starts. Your agent is working before you look away.

Trigger from anywhere.

Kick off agent tasks from the tools your team already uses. Results flow back automatically.

Slack

/caged run from any channel

GitHub

Label issue → agent starts

Linear

Assign to agent → sandbox

Jira

Move to Agent Review status

Discord

/caged run bot command

Scheduled Runs

Run agents on a cron schedule. Nightly test suites, weekly dependency audits, daily code reviews — all automated.

Webhook Triggers

POST to a webhook URL and Caged spins up a sandbox. Wire it to anything — Zapier, n8n, custom backends.

Results Everywhere

Agent posts results back to the originating thread, issue, or PR. Status, cost, trust score — all reported automatically.

The only sandbox with observability built in.

Other tools give agents a place to run. Caged tells you what they did, what they spent, and whether to trust them.

Cagedboxes.devE2BCodespacesLocal
VM-level isolation
Real-time cost tracking
Session replay
Trust scoring
Budget auto-kill
Network policy
Policy engine (RBAC guardrails)
Human-in-the-loop approvals
Agent eval & testing
Any agent (Claude, Cursor, Aider)
Sub-second boot
Snapshot & fork
CLI + API + Dashboard

Start free. Pay when you scale.

Use your own LLM API keys — we don't sell tokens. You pay only for compute and platform features.

Free

$0

Experiment freely

  • 5 sandboxes / day
  • 1 concurrent
  • 30-min timeout
  • 7-day history
  • 1 API key
  • Community support
Start free

Pro

$29/mo

Ship with AI daily

  • Unlimited sandboxes
  • 5 concurrent
  • 4-hour timeout
  • 90-day history
  • Unlimited keys
  • Budget guards
  • Priority support
Start Pro trial

Team

$99/mo

Agents at scale

  • Everything in Pro
  • 15 concurrent (shared)
  • Shared dashboards
  • Team key mgmt
  • SSO / SAML
  • Dedicated support
Contact us

Uses your existing Codex / Claude Code / OpenAI subscription — we don't sell tokens.

AI agents for
the whole team.

Your team is already using Claude Code and Cursor. Caged gives them isolated environments and gives you the controls to roll it out responsibly.

SSO / SAML / OIDC
SCIM provisioning
Role-based access
Audit logs
Spend controls
Network allowlists
Team dashboards
Budget per developer
Trust score alerts
Session replay sharing
Dedicated support
Custom MSA & DPA

FAQ

Still have a question? Get in touch.

How is this different from running agents locally?+

Locally, agents run with your credentials, your filesystem, and no guardrails. Caged gives each agent its own VM with network isolation, budget limits, and full observability. If something goes wrong, it's contained.

Can I use my existing Claude Code / Codex subscription?+

Yes. Caged doesn't proxy or resell LLM tokens. Your agent connects to Anthropic, OpenAI, or any provider using your own API keys. We track the estimated cost by observing token counts and model usage.

What agents are supported?+

Any agent that works over SSH or MCP: Claude Code, Cursor, Aider, OpenHands, Codex, or custom MCP servers. If it can connect to a Linux machine, it works with Caged.

How does cost tracking work if I bring my own API keys?+

The observability layer inside each sandbox intercepts LLM request metadata (model, token counts) — never your API keys. We estimate cost using a maintained pricing table of model rates.

What does 'trust scoring' actually measure?+

Risk signals: sudo attempts (-20), unknown network calls (-15), .env access (-25). Positive signals: running tests (+3), git operations (+5). The score starts at 100 and adjusts in real-time. You set the threshold for alerts or auto-kill.

Is my code safe?+

Your code runs inside a Firecracker microVM with its own Linux kernel. It never touches our host filesystem. Sandboxes are destroyed after use. We don't store your code beyond the session.

How fast do sandboxes boot?+

Under 300 milliseconds. We use Firecracker microVMs with pre-warmed pools — your agent gets a running VM almost instantly.

Can I self-host Caged?+

Self-hosted deployment is on our roadmap. The sandbox agent and MCP server will be open-sourced. Contact us if you need on-prem before the public release.

Ship faster.
Sleep better.

Your agents get their own environment. You get full control.

Free. No credit card. 30 seconds to start.