Bring artificial intelligence into your business without giving up control of your data. Powerful models, controlled infrastructure, and a solution built around the way you already work.
Estimate workload and token-based costs with your own assumptions. Envyon’s proposal depends on the configuration and contracted scope.
The initial €20/1M input and output rates preserve the old illustrative assumption; they are not provider prices. Replace them with your quoted rates. The initial 75% input share is a configurable example. Employee count estimates usage; it does not determine the required hardware.
ESTIMATED MONTHLY LOAD176M
token-equivalents across 88,000 interactions
Input: 132M · Output: 44M
22 working days per month
USAGE-BASED SERVICE€3,520estimated per monthENVYON · MANAGED INFRASTRUCTURE / RENTALProject-based quoteConfiguration and scope to be defined
This estimates token consumption only. It does not calculate savings against an Envyon quote or guarantee equivalent capacity. Hardware purchase is quoted separately.
Included in the calculation: input and output token consumption only. Excluded: VAT, setup, integration, energy, hosting, network, storage, embeddings, OCR, backups, support and internal staff time. The quote must specify included services. Capacity requires model, context, concurrency and latency tests.
Indicative model, not a quote. No per-token billing does not mean infinite technical capacity: the model, hardware, integration, support, and final capacity depend on the project.
Private RAG in practice
Private RAG: answers grounded in your company’s own knowledge.
Documents
→
Retrieval
→
Private model
→
Cited answer
A private RAG system connects an AI model to your approved documents so people can find grounded answers without exposing the knowledge base through a public service.
Hektorius is an autonomous robot with a personality tuned to yours. It learns how you express yourself so when friends drop by, it answers the way you would.