Replit
AI requests can land at 10-20x raw model cost. Fast, polished, and expensive when the agent runs hot.
☠ </> zero markup cloud IDE
A cloud IDE that passes AI costs through at zero markup. No platform tax. No vendor lock-in. Just build.
// bring your own keys
const ide = new PirateDev({
aiMarkup: 0,
lockIn: false,
deployTarget: 'your-vps'
})
await ide.agent.build(
'ship the thing',
{ provider: 'your-choice' }
)
The Problem
AI requests can land at 10-20x raw model cost. Fast, polished, and expensive when the agent runs hot.
Great if Azure is your harbor. Less great when your workflow needs to leave Microsoft's map.
Twenty bucks a month for model access you could wire directly, plus limits when the queue gets loud.
The Solution
Bring your own API keys. PirateDev passes model usage through at $0 markup, then charges only for compute. Your keys, your providers, your bill.
Actual API cost varies by provider, context, and output size. The markup stays dead.
Features
Monaco-based, dark theme, every language.
Claude, GPT, Groq — your keys, your choice.
Describe what you want. AI builds it step by step.
See changes instantly with hot reload.
Deploy to your own VPS, not our servers.
Full shell access. Run anything.
Open Source
AGPL frontend. Proprietary backend. See exactly what you're running. Fork it. Improve it. Just don't charge for it without a license.
Pricing
$0
Self-host, bring your own infra, community support.
$9/mo
Hosted IDE, managed VPS, priority support.
$29/seat/mo
Shared workspaces, team secrets, admin controls.
Get early access before the docks get crowded.