One endpoint in front of every model
An OpenAI-compatible endpoint on your domain, covering public providers and models on your own GPUs.
Infegate by DemirTech
Infegate is the OpenAI-compatible boundary in front of every model your organisation uses. Policy decides what may leave. The audit record proves what did.
01 / Production
Infegate runs in an unnamed customer environment and carries real workloads. The institution cannot be named. This statement does not claim a public reference, a contract state, or unverified performance figures.
02 / Boundary
Applications integrate once. Per-workload policy routes each request to an approved public provider or to open-weight models on GPUs inside the customer boundary. Identity, destination and cost are recorded at the crossing.
03 / Capabilities
An OpenAI-compatible endpoint on your domain, covering public providers and models on your own GPUs.
Prefix wildcards route new model identifiers without a configuration change.
Every key carries consumer metadata, so calls do not arrive as anonymous shared traffic.
Body logging is selectable per consumer, and the record keeps identity, model, destination, token counts and cost.
Usage is priced against a maintained rate catalogue for major providers.
A GitOps-friendly workload with an offline install bundle and a signed offline licence. There is no call-home check.
04 / Deployment
Air-gapped on the customer's Kubernetes cluster and GPUs. Regulated data does not leave. Open-weight models only.
Sensitive workloads remain inside while policy permits other traffic to approved public providers.
Public providers governed through one endpoint, one key model and one audit record.
05 / DemirTech
Infegate is a DemirTech product. Agents runs on top of the same boundary, but remains a separate ecosystem product rather than part of the Infegate offer.
Request a technical review