01
Send the request
Call one API instead of juggling several providers. Tell Dr Ney how much quality the task needs, or let your defaults decide.
Model routing API
Stop paying premium prices for easy questions.
Dr Ney sits between your app and your AI providers. Every request goes to the cheapest model that can do the job, and premium models are used only when the task needs them.
How it works
Three steps, no new infrastructure.
02
Dr Ney picks the model
Everyday questions go to fast, low-cost models. Only work that truly needs it reaches a premium model, and only if you allowed it.
03
Fall back, enforce, log
If a provider fails, Dr Ney switches to another. Your rules are enforced in code, and every call is logged.
What you get
Built to keep AI costs predictable.
One API, many providers
Groq, Cerebras, DeepSeek and Anthropic behind a single interface. Swap or add a provider without touching your app.
Cost-aware routing
Each request is matched to the cheapest model that can do the job, so you stop paying premium prices for easy questions.
Premium models locked by default
Expensive models stay off until you explicitly allow them. No surprise bills from a misrouted request.
Automatic failover
When a provider is down or slow, traffic moves to your next choice instead of failing in front of your users.
Guardrails per client
Define what each assistant may say and do. The limits live in code, not in a prompt that can be talked around.
Full trace log
See which model handled which request and why. Every agent action is recorded for review.
Savings
What routing is worth on a real workload.
Take 1,000,000 requests a month, each about 800 tokens in and 200 out. Here is the model bill, in US dollars as the providers charge it, with and without routing.
Everything on a premium model
$5,400
- Requests on premium100%
Routed by Dr Ney
$1,040 about 81% less
- Requests on fast models85%
- Requests on premium15%
Illustrative. Assumes premium pricing of $3 in and $15 out per million tokens, and fast-model pricing of $0.15 in and $0.75out. Your result depends on your traffic mix and on your providers' current prices, and it excludes the Dr Ney package fee.
Pricing
Simple monthly packages.
You pay each provider for the models directly. These packages cover the routing, failover and logging.
Starter
€29 / month
100,000 requests a month
€0.20 per extra 1,000
- Up to 3 providers
- 1 API key
- Automatic failover
- Email support
Most popular
Growth
€109 / month
1,000,000 requests a month
€0.12 per extra 1,000
- All providers
- 5 API keys
- Automatic failover
- Usage dashboard and trace log
- Priority email support
Scale
€349 / month
5,000,000 requests a month
€0.08 per extra 1,000
- Everything in Growth
- Custom routing rules
- Premium-model controls
- Unlimited API keys
- Support with a response-time target
FAQ
Questions we hear first.
Do I pay for the models through you?
No. You bring your own provider keys and pay each provider directly. Your Dr Ney package covers the routing, failover and logging.
How is the saving measured?
The example above uses stated assumptions. Your real saving depends on how many of your requests are simple. The trace log shows exactly which model handled each one, so you can check it against your own bill.
What happens if a provider goes down?
Dr Ney sends the request to the next provider you have configured, so your users keep getting answers.
Can I stop premium models from being used?
Yes. Premium models are off by default and are only used once you turn them on.
Ready to cut your model bill?
Tell us what you are running and we will set you up with an API key.