One balance. Every model. No overage bill, ever.
Start free with a daily message allowance and no credit card. Paid plans drop the daily cap and replace it with a monthly compute balance you can spend across any model in the library, in whatever mix you like — a thousand fast replies, or a much smaller number of very large ones.
Every number on this page is generated from the same plan catalogue the checkout reads, which is why you will not find a price written into the prose here. The table is the source of truth; the words around it are not allowed to disagree with it.
Free is a real account, not a trial that expires: a daily message cap, the smaller uncensored models, full chat history, no card. The first paid tier removes the daily cap and opens the full uncensored library. The tier above it adds the frontier models, Thinking mode, and higher context and output limits. The top tier is for people who live in these conversations all day and want the largest monthly allowance and priority routing.
| Plan | Monthly | Yearly | Message limit | Frontier models |
|---|---|---|---|---|
| Free | Free | — | 10 a day | No |
| Rogue | $10/month | $100/year | No daily cap | No |
| Pro | $20/month | $200/year | No daily cap | Yes |
| Max | $50/month | $500/year | No daily cap | Yes |
Paid plans include a monthly compute balance rather than a message count, because a message is not a unit of anything. A one-line reply from a small model and a long answer from a frontier model differ in real cost by more than an order of magnitude, and any product quoting you a flat message allowance is either overcharging the light users or quietly throttling the heavy ones.
The balance is visible before you spend it and while you spend it, so the tradeoff — cheap and fast, or expensive and deep — is yours to make deliberately rather than something you discover at the end of the month.
Rogue Compute is the unit paid plans are metered in: 10,000 Compute equals $1.00 of raw model cost. Metering the underlying cost rather than a message count means a short question to a small model costs a fraction of a long request to a large one.
You cannot be charged for more than you bought. When the monthly balance runs out, the account stops spending — there is no surprise invoice and no automatic escalation to a bigger plan. If you need more before the month resets, prepaid boosts add compute immediately and do not expire while the account is active.
| Pack | Price | Compute |
|---|---|---|
| Boost 10K | $4.99 | 10,000 |
| Boost 25K | $9.99 | 25,000 |
| Boost 50K | $19.99 | 50,000 |
It exists so you can answer one question before paying anything: does an uncensored model actually change your work, or does it just sound like it would? Ask it the thing you have been rephrasing for a filtered assistant. That takes about four messages, and the free allowance is sized for exactly that kind of evaluation rather than as a permanent free product.
If the answer is no, close the tab and you have lost nothing. If it is yes, the daily cap will start irritating you within a week, which is the honest signal that a paid tier is worth it.
Both billing periods are offered for every paid tier, with annual costing meaningfully less than twelve months of monthly; the exact saving is shown on the billing toggle and comes from the same catalogue as the prices. Checkout and card handling run through Stripe — OpenRogue never sees your card details.
Cancel from the billing portal whenever you like. Cancelling ends the subscription, not the account: your chat history stays where it is, still exportable as JSON and still deletable on demand from Settings.
A rough guide rather than a sales matrix, and it is fine to start lower than this suggests — upgrading takes a minute and the compute is what changes, not the product.
Free. Enough to establish whether the difference is real for your work.
The first paid tier. The full uncensored library and no daily cap, which is the whole complaint most people arrive with.
The tier that includes the frontier models and Thinking mode, because that is where explicit reasoning and the largest context windows live.
The top tier: the largest monthly allowance, the highest context and output limits, priority routing.
Yes — a daily message allowance on the uncensored models, with no credit card required. The exact cap and what each paid tier adds are in the plan table above, which is generated from the same catalogue the checkout uses so it cannot go stale.
It is the monthly balance paid plans include, spent across whichever models you use. Efficient models barely move the meter and the most powerful frontier models use considerably more, which is why plans are sized in compute rather than in messages — a message from a small model and one from a frontier model are not the same thing.
No. Spending stops when the balance is gone; there is no automatic overage charge and no forced upgrade. If you want more compute inside the same month you can buy a prepaid boost, which is optional and does not expire while your account is active.
Yes, from the billing portal, with no retention call and no cancellation fee. Your chat history remains in your account afterwards, and you can export it as JSON or delete it permanently from Settings at any point.
No. The uncensored models are the point of the product, not the upsell — the free tier runs them under a daily cap, and the first paid tier removes the cap and opens the rest of the uncensored library. Paying more is about the frontier models, Thinking mode and monthly capacity.
It costs less per month than paying monthly on every tier, and the saving is shown on the toggle above. The reason to pay monthly anyway is if you are still deciding — nothing about the product changes between the two, so there is no reason to commit early beyond price.