SaaS
- For whom
- A fast start without your own infrastructure
- Data location
- Shared EU infrastructure
- Effort
- Low
Hosting & pricing
Hosting & pricing: the right model for every requirement
Habicht runs the way your requirements demand – from ready in no time to full sovereignty in your own house. Choose the operating model first; with SaaS you pick one of four clear plans, and we bill the AI compute (inferencing) separately.
Operating models
Operating models at a glance
Dedicated Cloud
- For whom
- Hard compliance requirements
- Data location
- Fully separated infrastructure in the EU cloud
- Effort
- Low–medium
Hybrid
- For whom
- Sensitive data local, scaling from the cloud
- Data location
- Sensitive data local, compute in the EU cloud
- Effort
- Medium
On-premise
- For whom
- Highest sovereignty, regulated environments
- Data location
- Fully in your own house
- Effort
- Higher
The same workflows in every operating model – you'll see the concrete SaaS plans right below.
Request a demo & quoteSaaS plans
SaaS pricing: four clear plans
With SaaS on shared EU infrastructure you choose one of four plans. Need more? Move to a custom plan or to Dedicated Cloud, Hybrid or On-Premise.
| Starter | Team | Business | Enterprise | More | |
|---|---|---|---|---|---|
| Price | 130 €/ month | 250 €/ month | 700 €/ month | 2,000 €/ month | On request |
| Max. users | 10 | 50 | 200 | 500 | Custom |
| Storage | 2 GB | 5 GB | 50 GB | 200 GB | Custom |
| Endpoints | 1 | 3 | 5 | 20 | Custom |
| Assistants | 3 | 10 | 30 | 100 | Custom |
| Knowledge sources | 10 | 50 | 200 | 500 | Custom |
| MCP connections | 2 | 5 | 10 | 25 | Custom |
| Tools | 2 | 5 | 10 | 25 | Custom |
All prices net, plus VAT. · We bill the AI compute (inferencing) separately and transparently – see below.
Fair and transparent
We deliberately state the system and the AI compute (inferencing) separately. So it stays traceable what you pay for – with no hidden AI costs.
Inferencing
Inferencing: the AI compute – transparently separate
Whatever AI compute you use, we state it as its own line item – separate from the SaaS system. Two models are available.
Token packages
Prepaid credit packages (e.g. €50, €100 or €500). You only pay for the tokens you actually use – billed transparently per model.
Rent fixed compute
You rent fixed compute – predictable, with consistent response times, without counting tokens. A model that generic AI services don't offer this way.
Available with HabichtLLM.
- Standard
- up to 40 tokens/s per session
- PremiumOn request
- up to 200 tokens/s per session
Which AI models sit behind the inferencing – sovereign, European or optionally global – is covered under HabichtLLM.More on HabichtLLM
Dedicated cloud, hybrid & on-premise
More sovereignty when you need it
Dedicated Cloud
On request you get a fully separated infrastructure in the EU cloud – no shared resources with other customers. Sensible when compliance requirements demand strict tenant separation at the infrastructure level.
Scope and price on requestHybrid
Sensitive data and its processing stay local, compute-intensive tasks run in the EU cloud – controllable per use case. The dedicated-cloud option is available here too. You decide which models run locally or in the cloud.
Models: HabichtLLMScope and price on requestOn-premise
Habicht runs fully in your data centre – including a tailored setup, a suitable hardware option, support and maintenance from a single source. Local AI models are possible.
On-premise in detailScope and price on requestWhich AI models?
Sovereign, European – or global where it's uncritical
Habicht is multi-model: sovereign and European models for sensitive tasks, optionally global models for uncritical ones. You decide per use case.
More on HabichtLLMFAQ
Frequently asked questions
SaaS starts at €130/month (net) and spans four plans up to Enterprise; we bill the AI compute (inferencing) separately. Dedicated Cloud, Hybrid and On-Premise are quoted individually – best in a short demo.
The plan covers the platform with users up to the respective limit, storage and fixed quotas for endpoints, assistants, knowledge sources, MCP connections and tools. The AI compute (inferencing) is added separately and transparently.
A fully separated infrastructure in the EU cloud – without shared resources with other customers. Available from the Enterprise plan and in the hybrid model.
Instead of billing by tokens, you rent fixed compute per session – with predictable costs and consistent response times. Available with HabichtLLM in two tiers: Standard (up to 40 tokens/s per session) and Premium (up to 200 tokens/s per session).
Which model fits you?
In a short demo we find the right operating model and put together a quote for you.