Start small. Scale predictably.
A pricing ladder from pay-as-you-go compute to enterprise contracts. Most customers begin with a small RAG project and grow into dedicated endpoints.
Developer
Individual builders and small teams.
- Self-serve GPU / VPS
- Pay-per-token inference
- Shared inference endpoints
- Community docs & support
- Usage dashboard
Startup
AI-native startups shipping to production.
- Everything in Developer
- Training & fine-tuning jobs
- Dedicated inference endpoint
- Reserved GPU capacity
- Priority job queue
- Email support
Business
SMBs with company knowledge bases.
- Managed RAG assistant + API
- Data cleaning & evaluation
- Dedicated endpoint + SLA target
- Assisted onboarding
- Monthly plan + overage
- Business-hours support
Enterprise
Large companies and regulated users.
- Private / dedicated deployment
- SLA, SSO & audit logs
- Committed spend & reserved compute
- DPA + data deletion controls
- Named account manager
- 24/7 support
All plans include usage metering, project isolation, and the console. Prices in EUR, excl. VAT.
Every stream on one invoice
| Stream | Billing | Best for |
|---|---|---|
| Compute hosting | Hourly / monthly · storage · bandwidth | Developers, startups |
| Training / customization | Setup fee + job fee + add-ons | SMBs, agencies |
| Inference API | Per-token + plan + overage | All segments |
| Reserved capacity | Monthly reservation | Heavy users |
| Enterprise contracts | Annual + committed spend + SLA | Large companies |
| Professional services | Audit · integration · migration | SMB & enterprise |
Common questions
How does billing work?
A hybrid model: setup fee plus a monthly platform fee or minimum spend, plus usage-based inference. Enterprise moves to annual contracts and reserved capacity.
What's included in a RAG setup?
Data cleaning, embeddings, a vector database, the retrieval pipeline, a prompt template, an evaluation report, and a hosted OpenAI-compatible endpoint.
Can I bring my own model?
Yes — deploy open-source models or your fine-tuned artifacts to shared, dedicated, or reserved inference.
Do you offer a DPA and data deletion?
Enterprise plans include a DPA, deletion controls, encryption at rest, access logs, and private deployment options.
Make room for your next idea.
Bring your data, models, and compute into one workspace. Start with a project and build from there.