Public proof, live now

Serverless execution for stateful runtimes.

Real browser and sandbox microVMs on demand. 4 vCPU, 8 GB DDR5, root access, Docker, and NVMe — for $0.05/hr. Not a container. Not a restricted shell. A real machine, ready in under a second.

Start free Read docs

01 — Runtime

A real Linux
machine.

Not a restricted container with 1 GB of RAM. Every session gets 4 vCPU, 8 GB DDR5, root access, Docker, and a real kernel. The machine your agent runs on should not be smaller than the one you develop on.

Process tree — PID 1 owned

1 init (chaser)

ls /dev/kvm

crw-rw---- 1 root kvm 10, 232 ...

docker ps

CONTAINER ID IMAGE ...

0.46s Cold start to shell

8GB DDR5 5600 per VM

1.1GB/s Network throughput

14.7GB/s Disk I/O (mixed R/W)

02 — Browser

Invisible to
every detector.

Most browser automation is detectable because it patches JavaScript after the fact. Chaser's browsers are hardened below the runtime — consistent fingerprints, no injected artifacts, no leaked CDP traces. 16/16 detection tests passed. 21/21 real-world sites scraped — including Cloudflare-protected targets.

BrowserScan 82% authentic · no bot

CreepJS Unique fingerprint

Sannysoft All tests passed

CDP Detection Undetectable

WebRTC Leak No leak detected

Rebrowser Human-grade score

Tested daily against real targets

Google · Amazon · Cloudflare · Reddit · GitHub · eBay · Stack Overflow · Discord

21/21 targets · 100% success

// Connect with Playwright — nothing changes const browser = await chromium.connectOverCDP( `wss://${session_id}.chaser.sh` );

03 — Access

Your stack,
still yours.

Keep the editor, shell habits, and local Git workflow you already trust. Swap only the execution layer: local-first project mode, remote-first workspaces, MCP, REST, or SSH all land on the same real machine model.

DIRECT SSH ACCESS: ssh real-ssh@ssh.chaser.sh

curl -X POST https://api.chaser.sh/v1/sessions \
  -H "Authorization: Bearer $CHASER_API_KEY" \
  -d '{"session_type": "sandbox"}'

# → { "id": "ws_7f3a9c2e", "state": "running" }

04 — Workspace

Keep your code local.
Send execution away.

Local Project Mode keeps your checkout authoritative while a remote sandbox mirrors it for installs, tests, and agents. Durable workspaces hold the remote caches. The ledger gives you a clean rollback path when sandbox-originated edits come back home.

Local-first sync

Your editor and local files stay canonical. `/workspace` is the execution mirror.

Ledger + rollback

Sandbox-originated edits are recorded locally so you can inspect, undo, and recover with intent.

Remote-only caches

Heavy generated artifacts stay remote by default, so local machines stay clean while the sandbox stays warm.

MicroVM isolation

Every session still runs inside a real Linux microVM with the same hard isolation and tooling depth.

05 — Philosophy

Zero friction
by design.

One install. One command. Get access to infrastructure in seconds, and start for free.

Bootstrapping

Install in seconds.

One command to own your infrastructure.

Automation

Agent Configuration.

Instant, automatic MCP configuration across your stack.

06 — Comparison

You're overpaying
for less.

Prices from public pages. Performance from real benchmark runs inside each platform's default environment.

	Chaser Browser + Sandbox	Browserbase Browser	E2B Sandbox	Daytona Sandbox
Price	$0.05/hr 4 vCPU · 8 GB DDR5	$0.12/hr+ Overage rate · browser only	$150/mo Pro plan · 2 vCPU · 1.9 GB	$0.07/hr+ Per-resource billing · 1 vCPU · 1 GB
CPU	4 vCPU AMD Ryzen 9 9950X	Managed	2 vCPU Intel Xeon @ 2.6 GHz	1 vCPU Default tier
RAM	8 GB DDR5 5600 MT/s	Managed	1.9 GB	1 GB
CPU benchmark	3,166 Geekbench 6 single-core	N/A	1,405 Geekbench 6 single-core · 2.3x slower	N/A
Disk I/O	14.7 GB/s Mixed R/W · NVMe	N/A	2.7 GB/s 5.4x slower	N/A
Browser sessions
Sandbox + shell
SSH access
MCP support

Pricing sourced from Browserbase, E2B, and Daytona public pricing pages. CPU and disk benchmarks from YABS + Geekbench 6 run inside each platform's default environment.

Pricing

Pay for execution, not for babysitting machines.

Free gets you ephemeral access. Usage-based adds durable workspaces at $0.05 per session-hour. Subscription plans add account-level limits and team controls.

Free

For first runs, quick validation, and public exploration.

$0 /month

Ephemeral sessions only
1 concurrent session
1 hour session timeout

Start free

Usage-based

For automation workloads that need persistence without a seat fee.

Proof-first

$0.05 /session-hour

Metered per second
$10 minimum wallet top-up
Persistent workspaces enabled

Add credits

Subscription

For teams that want predictable limits, observability, and shared control.

Plan limits

Plan-specific concurrency caps
Plan-specific workspace and session limits
Persistent workspaces enabled
Usage metering stays visible for observability

Choose a plan

Billing is account-scoped. Personal and organization accounts can be priced differently.