Your infra,
run by Agents.

Deploy, scale, and secure your infrastructure using the Scale Agent, or inside Claude Code via MCP. Scale+ adds security scanning, load testing, and automated protection.

Scale Agentweb

Deploy, manage, and protect your infrastructure directly in the dashboard using the agent.

Open dashboard
Claude Code MCPcli
claude mcp add scale-platform --transport http ...
Open MCP docs
Get started free

Two ways to control your infra

Use the Scale Agent on the dashboard, or connect Claude Code via MCP.

Claude Codescale-platform
You›Deploy my-api from the main branch
● Calling create_project
✓ Project created → dep_a3f9
● Calling deploy_project
✓ Build started
You›Run a security scan and save a report
● Calling scan_security
✗ [high] Missing X-Frame-Options
✗ [medium] No Content-Security-Policy
✓ [pass] No XSS reflection found
● Calling report_finding
✓ Report saved
You›Load test it and verify autoscaling
● Calling verify_autoscaling
✓ Scaled 1 → 3 replicas under load
Scale Platform
Project
my-apiready
my-api.apps.scale.searchhq.org
replicas
3
cpu
12%
mem
128mb
Security Report
Security Audit — my-apihigh
2 findings · saved to dashboard
✗ Missing X-Frame-Options
✗ No Content-Security-Policy
Load Test
rps
842
p99
38ms
errors
0%
✓ Autoscaling triggered at 60% CPU
1234TypeError: undefined

Manage deployed services

Import a repo and trigger deployments from any branch.

Scale replicas

Scale any project with automatic load balancing.

THRESHOLD245ms

Run sandboxes

Create isolated environments and execute code instantly.

autonomous

Self-Driving Infrastructure

Scale watches your platform around the clock. When a threat or anomaly is detected, the agent investigates and acts — no pager, no intervention needed.

Trigger conditions
⬡
Honeypot hit
A request reached a trap path like /.env or /admin/config — attacker detected.
↑
CPU spike
Container CPU exceeds threshold — agent checks for runaway processes or load.
✕
Error spike
5xx error rate spikes above baseline — agent reviews logs and traces root cause.
↯
Deploy failure
A deployment fails — agent inspects build logs and surfaces the exact error.
Autonomous Agentlive
⬡honeypot hit · /.env · 203.0.113.47
→ querying 50 security events...
→ 12 hits from same IP in 15 min
✓ 203.0.113.47 blocked
✓ Slack notified
↑cpu spike · my-api · 94% for 3 min
→ checking container stats...
→ load spike, autoscaling at 80%
✓ scaled 2 → 4 replicas
✕error spike · api-gateway · 8.3% 5xx
→ reviewing container logs...
investigating…
K8sK8s
LinearLinear
BashBash
DockerDocker
GoGo
Node.jsNode.js
PythonPython
SlackSlack

Sandbox execution environments — ask the agent to create one and run code instantly.

Let agents manage your infra.

Sign up and use the Scale Agent in your browser, or connect via MCP to Claude Code.

Sign up with GitHubView docs