pricing
AI that gets it right, from €9.99 a month. Start free, no card.
Each optimization is a real call to a model and costs us money, so what you get is a number of them per month, written down and the same for everyone. No «fair use policy» waiting for you in the small print.
Free
0 € /forever
To see if it helps.
Anyone. No card, no expiry.
- 5 AI optimizations a month
- Marked gaps instead of invented details
- Word-by-word diff against your original
- Prompt analysis before you optimize
- 7-day history
- 3 custom dictionary rules
Pro
€6.67 /mo
€79.99 a year · you save €39.89
€0.22 a day
Everyday use, with AI.
Anyone working with prompts every day.
Ten euros more than Basic and the counter disappears, the API opens up and your history stops expiring.
- 150 AI optimizations a month
- Aggressive mode with AI (counts as 3 credits)
- Prompt quality scoring
- Optimization for your target model
- History kept forever
- Unlimited custom dictionary
- Everything in Free
Power
€13.33 /mo
€159.99 a year · you save €79.89
€0.44 a day
Volume and automation.
Anyone optimizing in bulk or wiring it into their own code.
- 400 AI optimizations a month
- Batch optimization
- More optimizations at once
- API to wire it into your own code
- Everything in Pro
Team
€33.33 /mo
€399.99 a year · you save €199.89
€1.11 a day
One standard for everyone.
A team or agency with several live projects.
Each seat costs a little over half a Pro licence. From three people up.
- 1,000 AI optimizations a month
- Up to 5 people
- Shared dictionaries
- Team analytics
- Role-based access control
- A single invoice
- Everything in Power
Prices exclude VAT, which is added at checkout based on your country. Cancel whenever you want from your account and you keep access until the end of the period you paid for.
Why we can take the counter away
Optimizing you costs us zero. The engine is rules and dictionaries running in your browser, not calls to a model. Try it in airplane mode: it keeps working.
That is why there is no asterisk. Anyone with an LLM behind it pays for your every use and is forced to meter you, even if they write «unlimited» and qualify it in the small print. We do not have that problem.
You pay us for what does cost: storing your history, your rules and your team's work. That is real servers. Optimizing is not.
Enterprise
Self-hosted engine inside your own infrastructure, SSO access, a service level agreement, a data processing agreement and unlimited seats. For when the prompts cannot leave your network.
Frequently asked questions
Tap any of them to open the answer.
How many optimizations does each plan include?
The free plan includes 5 AI optimizations a month, Pro 150, Power 400 and Team 1,000. The counter resets on the 1st of each month and is enforced on the server. Deep rebuild mode, which redoes the whole prompt, counts as three.
So what am I paying for?
Mostly volume: five a month is enough to try it, not to work with it. On top of that, Pro adds deep rebuild mode, quality scoring and permanent history; Power adds batch optimization and the API; and Team adds up to five people with shared dictionaries and analytics.
Can I cancel whenever I want?
Yes, from your account and without writing to anyone. You keep access until the end of the period you already paid for, and when it ends the account returns to the free plan: your rules and your history are not deleted, the free plan limits simply apply again.
Is this not the same as asking ChatGPT to improve my prompt?
It is similar, and the difference is one rule: making things up is forbidden here. If you did not state the length, the tone or the recipient, the result comes back with a marked gap in brackets rather than a decision the model took for you. Asking for it directly usually returns a prompt with requirements you never set, and you cannot tell which ones.
Does the prompt come out shorter?
Usually not: it tends to come out longer, and that is the point. What gets optimized is the clarity of what you are asking for, not its size. A nine-word prompt can become forty if it was missing the goal, the output format and the data. You see the token count before and after, with its real sign.
Does optimizing use up tokens from my ChatGPT or Claude account?
No. We pay for the rewriting from our own Anthropic account; it costs you one credit from your Prampt plan. What you do afterwards with the resulting prompt is on your own account, like any other message you send.
Can it change what my prompt means?
The engine is forbidden from removing or changing what you did say, and in deep rebuild mode a second model scores the result and corrects it if it falls short. Even so it is written by a model, so every result comes with a word-by-word diff: the decision to use it is yours, with the evidence in front of you.
Do you store my prompts?
No. What each optimization cost and how long it took is recorded, never the text. From your history we store metrics and a 120-character preview so you can recognise each entry. The prompt is sent to Anthropic to be rewritten, and it is not used there to train models.