Skip to content

Vex-Hub integration

All of Vex-Hub's LLM trafficruns through Vex Base.

Applications on Vex-Hub never connect to model providers directly. Every call passes through the Vex Base gateway, permissions, budgets and records are defined in one place, and control stays with your organisation.

One gateway

Vex-Hub Applications Go Through the Same Door

From the field application to the customer portal, every Vex-Hub application makes its model calls through Vex Base. That leaves one path to the models across the organisation — and you write the rules for that path.

Vex-Hub · model traffic

Connected to Vex Base

Vex-Hub

3 applications
Field applicationCustomer portalInternal dashboard

Vex Base gateway

One policy

Permissions

RBAC

Budget

$6,500/mo

Records

All calls

GPT-5.1
Claude Opus 5
Gemini 3 Pro

Permissions and budget

You Decide Which Application Uses Which Model

For each Vex-Hub application, define the models and data sources it may reach and its monthly limit. Permission changes take effect immediately, and once the limit is hit the application falls back to a cheaper model.

Access matrix

+ New role

Head of Legal

Full access

General Assistant

Limited access

CFO Assistant

Admins only

Code Reviewer

API access

Guest

Read only

One invoice

Vex-Hub Consumption Sits in the Same Panel

Calls, tokens and costs from the Vex-Hub side merge with your other internal applications. You see what each application spends in one panel and receive a single invoice.

Governance › API consumption by application

Use caseCallsCost
Customer Support Portal18.6K$61.40
Contract Management9.4K$34.70
Field Service Application7.8K$28.10
HR Assistant4.3K$16.90
Developer Platform6.1K$25.30

Uygulama

5

Total calls

46.2K

Single invoice

$166.40

Enterprise security

Every Call on Record, Traffic Encrypted End to End.

Requests arriving through Vex-Hub pass along the same security chain: traffic is encrypted end to end, calls are written to records that cannot be altered afterwards, and the data never leaves the corporate perimeter.

The new standard for enterprise AI

Run artificial intelligenceon your own infrastructure.