Partner program
Your AI product. Our infrastructure.
For software agencies that build AI features for their clients — and want recurring infrastructure revenue without building an infrastructure team. You own the client and the product. We run the private AI layer underneath.
Partner architecture
- Your agency
- Your client's application
- White-label endpoint
- Lirux infrastructure
Who it's for
Software agencies whose clients are asking for AI.
The programme is designed for agencies of roughly 10 to 100 people, wherever they are based. You build web applications, SaaS, e-commerce or internal tools — and your clients want chatbots, search, document processing or transcription inside them.
- You deliver client projects and want AI features to be a standard part of your offer
- Your clients ask where their data is processed and expect a clear answer
- You prefer not to run GPUs, inference servers or model updates yourself
- You want to quote AI infrastructure as a fixed monthly line item
What partners get
Infrastructure, pricing and engineers behind your projects.
Partner pricing
Private nodes
Client-specific environments
Branded API domains
Deployment assistance
Centralised infrastructure
Engineering support
How the partnership works
From first conversation to your first client in production.
- 1
Apply
Tell us about your agency, the clients you serve and the AI features they are asking for.
- 2
Introductory call
A technical and commercial conversation: typical workloads, data requirements and how you want to package AI for clients.
- 3
Partner agreement
Partner pricing, the data processing agreement between us and you, and support arrangements — agreed per partner.
- 4
First client deployment
We deploy the first client environment together with your developers and confirm the integration works end to end.
- 5
Scale with your client base
Add client environments, nodes and models as projects come in. Capacity for larger expansions is planned with you in advance.
Business model
You bill your client. We bill you.
AI features need infrastructure for as long as the feature is live. That turns a one-off project into an ongoing service you can charge for every month — while the operational work stays with us.
- Lirux
- Your agency
- Your client
- We invoice you
- Monthly, in EUR, at your agreed partner pricing for the capacity your clients use.
- You invoice your client
- Under your own terms and pricing — as part of a retainer, a managed service or a per-feature fee.
- Recurring, not one-off
- Infrastructure and operations become a monthly line item that runs as long as the client uses the feature.
- No infrastructure overhead
- No hardware, no on-call rota for inference servers, no model maintenance on your side.
Clear responsibilities
What we provide. What stays yours.
Lirux provides
- Hardware, networking and hosting
- Model deployment, quantization and updates
- OpenAI-compatible endpoints per client
- Monitoring, alerting and runtime maintenance
- Technical support for your engineers
- Architecture advice for client projects
Your agency owns
- The client relationship and contract
- The client application and its user experience
- Pricing and packaging towards your client
- First-line (tier 1) support for your client
- Deciding which data the application sends
FAQ
Partner questions.
Partner terms are agreed individually. These answers describe how we usually work.
How does the data processing chain work?
Typically your client is the controller, your agency acts as processor, and Lirux acts as your sub-processor under a data processing agreement between us and you. Your client must be informed about sub-processors, including that compute currently runs in Georgia, outside the European Economic Area. Where EEA personal data is involved, transfer safeguards such as Standard Contractual Clauses may be required. Customers remain responsible for determining the appropriate legal basis for their workloads. This is not legal advice.
What does white-label cover — and what not?
White-label covers the technical interface: endpoints under a domain you control where technically available, and client-facing naming in your own application. It does not change the legal picture: where personal data is processed, we remain a sub-processor and must be disclosed as such, together with the processing location.
Are there minimum commitments?
Minimum volumes and terms are discussed per partner. We do not publish a partner price list; agency pricing depends on the mix of shared and dedicated capacity your clients need.
Can your engineers join client calls?
Yes, where it helps. Our engineers can join technical conversations with your client under your lead — the client relationship stays with you. Any joint sales arrangements are agreed per partner.
How does technical support work?
Your team provides first-line support to your clients. We support your engineers on infrastructure, model deployment and endpoint issues, in English and German. Support channels and response targets are agreed per contract.
Does every client need a dedicated node?
No. Smaller projects can run on the shared Managed AI API, while clients with sensitive data or steady volume get their own private node. Because the API is the same, a client can move from shared to dedicated without changes to the application.
Apply
Apply for partner access.
Tell us about your agency and your clients. An engineer reviews every application and replies by email to arrange an introductory call.
Prefer email? [email protected]
Offer AI to your clients without running AI infrastructure.
Start with one client project. We deploy it with your developers and you decide how to package it.