One engine on every plan.
Only the limits change.
Every plan ships the full grades, the citations and the refusals. What changes is how many requests a key can make per minute and per day, and how many units it can use per month. One answer is one unit; a protocol generated in chat or a discovery run is two, and a book is eight.
Free
Free
Every account in the iOS app.
- 20 requests a minute
- 300 requests a day
- 3,000 units a month
- A grade and its citations on every answer
- A refusal when the evidence is too thin
- Contraindication and interaction checks
Starter
Free
What every new portal account starts on.
- 60 requests a minute
- 5,000 requests a day
- 5,000 units a month
- A grade and its citations on every answer
- A refusal when the evidence is too thin
- Contraindication and interaction checks
Pro
Talk to us
Higher throughput for production integrations.
- 300 requests a minute
- 50,000 requests a day
- 200,000 units a month
- Everything in Starter
Scale
Talk to us
High volume and enterprise throughput.
- 1,200 requests a minute
- 500,000 requests a day
- 5,000,000 units a month
- Everything in Pro
Every plan, side by side
The same engine on every plan. The grades, the citations, the refusals and the safety checks come with all of them. Limits are the only thing you are buying more of.
| Feature | FreeFree | StarterFree | ProTalk to us | ScaleTalk to us |
|---|---|---|---|---|
| Limits per key | ||||
| Requests per minute | 20 | 60 | 300 | 1,200 |
| Requests per day | 300 | 5,000 | 50,000 | 500,000 |
| Units per month | 3,000 | 5,000 | 200,000 | 5,000,000 |
| Where | iOS app | Portal and API | Portal and API | Portal and API |
| Price | Free | Free | Talk to us | Talk to us |
| Evidence | ||||
| A grade on every answer | Yes | Yes | Yes | Yes |
| Citations to primary sources | Yes | Yes | Yes | Yes |
| Refusal when the evidence is thin | Yes | Yes | Yes | Yes |
| Safety and hosting | ||||
| Contraindication and interaction checks | Yes | Yes | Yes | Yes |
| An audit record of every answer | Yes | Yes | Yes | Yes |
| Servers and backups in Germany1 | Yes | Yes | Yes | Yes |
| Free of advertising, everywhere | Yes | Yes | Yes | Yes |
- 1Data and encrypted backups are kept in Germany. The AI models that write the answers run with providers in the United States. Where data is processed
Questions
What do the plans cost?
Free and Starter cost nothing. Pro and Scale are priced with you, so talk to us. There is no card to enter and no payment taken through the portal today.
What counts against a limit?
Per minute and per day, every API request made with a key. Per calendar month, units: one per answer, two for a protocol generated in chat or a discovery run, eight for a book. An organisation can also have a monthly cap across all its keys. When a window runs out the engine answers 429 with the time until it resets.
Can we move between plans?
Yes. Start on Starter and move up as traffic grows. The limits are set per key and change when your plan does.
Is any of it paid for by advertising?
No. No commercial interest belongs inside a prescribing decision, so nothing in the product is sponsored, promoted or placed.
What if we want the engine in our own software?
That is the engine licence. It is metered per call and it is sold to practice software, ERP and CRM vendors. Tell us what you are building and we will size it.
Put the engine inside your own software
Practice software, ERP and CRM vendors licence the engine directly. It is metered per call, and it carries the same grades, the same citations and the same refusals.