Plans & limits

Every Claudin.io plan is unlimited usage with a spend-protection cap. You're not billed per token or per request — you pay a flat monthly price and use it freely. The cap exists only to stop a runaway agent (an infinite tool loop, for example) from draining your plan.

The plans

Plan	Price	Spend protection	Best for
Starter	$5 / mo	$0.50 / hour	Trying out — low commitment
Lite	$9 / mo	$1.00 / hour	Hobby projects, occasional coding
Essential	$19 / mo or $189 / yr	$2.00 / hour	Quality for everyday use
Pro ★	$39 / mo or $389 / yr	$4.00 / hour	Heavy agentic workflows
Power	$59 / mo or $589 / yr	$6.00 / hour	Teams, multiple projects
Ultra	$99 / mo or $989 / yr	$10.00 / hour	Maximum power, teams & production

Most people never hit the cap

The hourly cap is generous for normal interactive work. You typically only brush against it if an agent goes into a tight loop — which is exactly when you want a brake.

Which model should you choose? Claudinio vs Claudius

We offer two main models for your coding agent:

Model	Backend	Use case	Recommended for
claudinio 🏆	Fast, balanced, cost-effective	Everyday coding, hobby projects, general code	All plans (Starter to Ultra)
claudius ★	Premium, deep reasoning	Complex tasks, deep reasoning, heavy agentic workflows	All plans (Starter to Ultra)

Straight talk

If you're on Starter ($5) or Lite ($9) — use claudinio and don't look back. 🎯

Here's the reality: claudinio delivers quality comparable to Claude Sonnet for everyday coding at a fraction of the internal cost. On the Lite plan, you can get hundreds of requests per hour with claudinio — while claudius would burn through your hourly budget much faster.

Metric	claudinio	claudius
Impact on hourly budget	Low — stretches much further	High — 6x per request
Use case	Daily coding, personal projects	Heavy reasoning, complex agents

Golden rule: Configure your agent (Claude Code, Cursor, Continue, etc.) with claudinio as the default model. Only switch to claudius when you explicitly need more reasoning power. For hobby projects, claudinio is all you need and probably more than you expect.

💡 Tip: Both models work with all major coding agents. Just set model=claudinio or model=claudius in your agent config. claudinio also automatically resolves aliases like claude-sonnet-4, gpt-4o, o3-mini and dozens more — no need to change your agent's configuration.

How spend protection works

Each plan defines a budget window — a rolling period and a maximum spend inside it:

• Starter, Lite, Essential, Pro, Power, and Ultra use a 1-hour window.

Within the window, your usage accumulates a tiny internal cost. When that internal cost reaches the window's cap, requests pause until the window resets.

Only your model calls through the proxy. Each request adds to the current window's running total based on the tokens it used. When the window resets, the total resets with it.

If you hit the cap and get a budget error, you have two options:

1. Wait for the window to reset (shown in your dashboard).
2. Upgrade to a higher plan for a larger cap.

See Plans-related errors for what the budget error looks like.