Plans and limits
What's free, what Pro unlocks, how the monthly AI usage bar works, and how to pick between Claude Sonnet and Claude Opus.
Everything in the manual editor — timeline, effects, captions imports, export — is free with no caps. What's metered on the free plan is the AI: every month you get a free AI budget that covers agent editing and AI video analysis, and when you want more than that, Pro removes the ceiling.
Free vs. Pro
| Free | Pro | |
|---|---|---|
| Manual editing and export | Unlimited | Unlimited |
| AI agent edits | Monthly AI budget | Unlimited |
| AI video analysis (the agent watching your footage) | Shares the same monthly budget | Unlimited |
| Model picker (Claude Sonnet / Claude Opus) | Included | Included |
| Price | $0 | $20/month or $99/year |
Upgrade at pluged.ai/pricing.
The AI usage bar
On the free plan, a thin AI usage bar sits at the top of the agent panel, and the same meter appears inside the model picker popover (with "% used" and "% left this month"). It shows one number: the percentage of your monthly free budget you've spent. Raw token counts are never shown — the percentage is the whole story.
The budget covers the two things that run paid AI inference:
- Agent editing — every message you send to the agent and the work it does in response.
- AI video analysis — when the agent watches your footage or reviews the rendered cut. See /docs/agent/visual-analysis.
Manual editing, playback, undo, and export never touch the budget — they run on your machine.
The bar refreshes after each agent turn settles. On Pro, there's no bar at all: the picker just shows "Unlimited".
What happens at 100%
When the bar fills:
- New agent requests are blocked. Sending a message returns "You've reached your free AI usage for this month," and the panel shows an Upgrade button (which opens pluged.ai/pricing) alongside Dismiss.
- A run already in progress finishes. The limit is enforced when a new request starts, so the agent won't stop mid-edit and leave your timeline half-done.
- Everything non-AI keeps working. You can keep editing manually and export your project as usual.
The budget is a monthly allowance, so it resets each month.
Choosing a model
The model picker is the settings control at the bottom-left of the agent composer. Two options:
- Claude Sonnet (default) — fast, and the right choice for almost every edit.
- Claude Opus — a heavier model for edits that need more judgment: dense multi-step briefs, tricky editorial calls, edits Sonnet didn't nail.
A practical pattern: stay on Sonnet, and if a hard edit comes back not quite right, switch to Opus and retry the same prompt. The same popover also holds the Review before applying toggle — see /docs/agent/how-it-works.
Limits
- The usage meter requires being signed in — the budget is tracked per account, not per project.
- The percentage is intentionally the only number shown; there's no token-level breakdown to inspect.
- Blocked means blocked for AI only: no agent runs and no new AI video analysis until the month resets or you upgrade.