Multi-model routing
Send each request to the best model for quality, speed and cost, with automatic failover.
Give employees and applications safe, controlled access to leading AI models, with full cost transparency.
Send each request to the best model for quality, speed and cost, with automatic failover.
Permissions by department, project and data sensitivity, integrated with enterprise SSO.
Automatically redact personal data, block prohibited content and stop prompt injection before data leaves.
Track spend by user, team and application, set budgets and get alerts on overruns.
Record every request and response for audit, investigation and compliance reporting.
Run on in-country cloud in Vietnam or on-premise, alongside self-hosted open models.

Connect to leading commercial and open-source models through one API. When a better model arrives, just change the configuration.
When every department adopts its own AI tools, the enterprise loses control of data and cost.
Sensitive data is redacted before it reaches any model.
Staff get an official, safe tool instead of personal accounts.
Know exactly who uses what, how much and with what value.
Full evidence for internal audit and regulators.
Popular commercial models and open-source models, cloud-hosted or self-hosted. The list is updated continuously.
Added latency is tiny compared with model processing time, and caching can reduce total response time.
Yes. AI Gateway can be deployed standalone or together with the full VDX AI OS.