Sovereign AI agents,
matching your ambitions.
Two products, two pricing logics. Tell us what you are looking for and we will show you yours.
The two combine, and that is where it gets interesting — see how ↓
To try it out, or for a single workstation that shares nothing.
- 1 user, on your own machine
- No sharing between users
- Open models pre-loaded
- No account, no API key
- No business connectors, no logging
- Community support
For a team that works with its own data.
- Multi-user, three-role governance
- Connectors to your business software
- Logging and processing record
- AgentHub console included
- Updates, fixes and maintained models
Any one of these four points is enough for us to build the offer with you.
- Several companies or several instances
- A service contract with penalties
- An agent that writes into your ERP
- A security audit or questionnaire
Three ways to deploy it, at the same price per user.
- Individual position. On each person's machine. Request a quote
- Dedicated server. Power shared by the team, either at your premises or in a Swiss datacenter. Request a quote
- Dedicated instance. We host and maintain it in Switzerland. Request a quote
Annual billing, per user. The Business licence is purchased in a block of ten seats, whether you are three or ten. Support and updates are included. The hardware stays in your name, at the price of the day, or you use your existing machines. Setting up is quoted after a twenty-minute scoping call, with no commitment. See WivenLLM in detail →
The console, the catalogue, and the trace of what your agents do.
- Control and supervision console
- Catalogue of agents ready to deploy
- Processing record and export
- Governance and roles
- Swiss hosting, or on your own premises
You pay for the agents you deploy. No more, no less.
- No limit on the number of agents
- Agents from the catalogue, or built for you
- Set-up and integration quoted on request
- Training for your teams included
- External resources passed on at cost, with no margin
Any one of these four points is enough for us to build the offer with you.
- Several companies or several instances
- A service contract with penalties
- An agent that writes into your ERP
- A security audit or questionnaire
One more agent can be added at any time, without renegotiating your contract. Setting up each agent is quoted after a scoping call. On a WivenLLM deployment, the external resources line comes to zero. Discover AgentHub in detail →
One principle, and it fits in two sentences.
Whatever runs on your own hardware is fixed-price. Whatever consumes a resource we buy on the outside is passed on to you at the price we pay for it, within a cap agreed with you at signature.
We take no margin on that usage. That is deliberate: it means we have no interest in your agent being talkative.
How does consumption work?
Your agents sometimes consume resources that we buy on the outside, such as a call to a third-party model, image generation, a web search or a data enrichment. We pass these on at the price we pay for them, with no margin, with a detailed monthly statement. We estimate the amount with you at the scoping stage and you set a cap.
This line only appears if your agents actually call an external resource. Many deployments have none: on a fully local WivenLLM deployment, it is zero.
What the set-up fee covers
Support is included. What we quote for is what we build.
We define together, at the scoping stage, the scope of each agent: what it does, on which data, with which tools. That scope is written down, it appears in your contract, and it is what answers every billing question.
We adjust an agent inside its scope.
- Your questions about using it
- Settings and fine-tuning of an existing agent
- Fixes and updates
- Model maintenance
- A broken agent is repaired
We move outside the scope, so we build something new.
- An additional agent
- A new business tool connected
- A process taken over or extended
- A new data source
- Quoted as a set-up
If your process cannot stop and you need a contractual availability commitment, that falls under the Enterprise offer and is built with you.
The agents decide what to do. WivenLLM decides where it happens.
An AgentHub agent can run on our platform, or on your own WivenLLM. The work is the same. What changes is your invoice.
| Line | On our platform | On your WivenLLM |
|---|---|---|
| Platform | Included | Included |
| Agents Verification Status | CHF 49 per agent, per month | CHF 49 per agent, per month |
| External resources | Passed on at cost, with no margin | CHF 0.– |
The computation happens on your own hardware, so there is nothing to buy on the outside, so there is nothing to pass on. That is the only difference, and it is the only one that shows on an invoice.
WivenLLM lets you define which requests run on your local model and which may call an external one, workspace by workspace. By setting that rule, you set your spending ceiling.
What people ask us before signing.
What is the difference between WivenLLM and AgentHub?
Do I pay per request?
How much does WivenLLM cost?
Does the price increase with each new hire?
What does the free version contain?
Is support included?
What exactly does the set-up fee cover?
Do we have to replace our hardware?
What happens if we want to leave?
Let's talk about your project.
Twenty minutes to understand what you need and give you a figure. If there is no measurable gain, we say so.
Talk to us — 20 min →