
Sep 11, 2026
IN /
Qantara
1 min read
Qantara: AI Gateway control plane complete and verified end to end
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.

IOTA Labs Team
Data & AI Practice
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.
The Qantara AI Gateway control plane is implemented: adapters for vLLM, Azure, AWS and GCP backends, a PostgreSQL spend ledger with catalog-priced usage, inline policy enforcement, an OIDC-protected admin API with an embedded console, and OpenTelemetry audit, all packaged as a Helm chart.
The acceptance test stands up the whole platform on a cluster and asserts the behaviours that matter: completions route by model name, usage lands in the ledger, an exhausted budget returns 429, a disallowed model returns 403, and the registry serves the curated catalog plus auto-discovered servers.

Sep 11, 2026
IN /
Qantara
1 min read
Qantara: AI Gateway control plane complete and verified end to end
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.

IOTA Labs Team
Data & AI Practice
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.
The Qantara AI Gateway control plane is implemented: adapters for vLLM, Azure, AWS and GCP backends, a PostgreSQL spend ledger with catalog-priced usage, inline policy enforcement, an OIDC-protected admin API with an embedded console, and OpenTelemetry audit, all packaged as a Helm chart.
The acceptance test stands up the whole platform on a cluster and asserts the behaviours that matter: completions route by model name, usage lands in the ledger, an exhausted budget returns 429, a disallowed model returns 403, and the registry serves the curated catalog plus auto-discovered servers.

Sep 11, 2026
IN /
Qantara
1 min read
Qantara: AI Gateway control plane complete and verified end to end
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.

IOTA Labs Team
Data & AI Practice
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.
The Qantara AI Gateway control plane is implemented: adapters for vLLM, Azure, AWS and GCP backends, a PostgreSQL spend ledger with catalog-priced usage, inline policy enforcement, an OIDC-protected admin API with an embedded console, and OpenTelemetry audit, all packaged as a Helm chart.
The acceptance test stands up the whole platform on a cluster and asserts the behaviours that matter: completions route by model name, usage lands in the ledger, an exhausted budget returns 429, a disallowed model returns 403, and the registry serves the curated catalog plus auto-discovered servers.

Sep 11, 2026
IN /
Qantara
1 min read
Qantara: AI Gateway control plane complete and verified end to end
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.

IOTA Labs Team
Data & AI Practice
Model routing, budgets, a spend ledger and audit now run through the Go control plane, exercised by an automated acceptance test on a Kubernetes cluster.
The Qantara AI Gateway control plane is implemented: adapters for vLLM, Azure, AWS and GCP backends, a PostgreSQL spend ledger with catalog-priced usage, inline policy enforcement, an OIDC-protected admin API with an embedded console, and OpenTelemetry audit, all packaged as a Helm chart.
The acceptance test stands up the whole platform on a cluster and asserts the behaviours that matter: completions route by model name, usage lands in the ledger, an exhausted budget returns 429, a disallowed model returns 403, and the registry serves the curated catalog plus auto-discovered servers.
(iota โ 11)
Product updates
More updates
More updates
More updates
Notes on AI systems, architecture decisions,
and lessons from real deployments.
No hype. Just systems
Clarity beats automation
Decisions over demos
Designed for messy reality
Systems that hold under pressure

(iota โ 15)
OUR PRINCIPLES
Book a
free call
We design AI systems that improve real work โ
not just demonstrate technology.
We design AI systems that improve
real work โ not just demonstrate technology.
Weโll review your
workflows, identify
where AI can create
impact, and outline
a clear path forward.
Weโll review your workflows, identify AI opportunities, and outline a clear path forward.
Weโll review your workflows, identify where AI can create impact, and outline
a clear path forward.
No preparation needed โ weโll guide the conversation
and focus on what matters.
No preparation needed โ weโll guide the conversation and focus on what matters.
7 AI products
100%
deployable on-prem, air-gapped




