witkeepRequest a demo ↗

Architecture

An AI gateway to control requests to models and tools. A business workspace for your teams.

Explore the three configurations ↓

API gateway / AI gateway

API gateway

Manages access to APIs across your information system.

AI gateway

Manages model requests, tokens and AI costs.

Witkeep in your architecture

Your setupWitkeep adds
No gateway
Witkeep adds

An AI gateway and the business workspace

An API gateway
Witkeep adds

AI request management and the business workspace

An AI gateway
Witkeep adds

The business workspace, connected to your gateway

The business workspace: contexts, skills, shared deliverables and ROI measurement.

Technical capabilities

Models & routing
Multiple providers, fallback and cost / performance / availability trade-offs.
Access & contexts
One unified Witkeep key. Permissions and restrictions per context.
Data protection
Guardrails, personal data masking and budget quotas.
MCP Gateway
Authentication, permissions and logging for tool calls.
Observability
Prometheus and OpenTelemetry for centralised monitoring.
Deployment
Dedicated cloud instance or on-premises, on Kubernetes.

Your architecture questions

Read the full FAQ ↗
What is an API gateway?

An API gateway is an entry point between applications and services. It lets you control access, apply rules to requests and monitor how those exchanges work.

What is the difference between an API gateway and an AI gateway?

An API gateway controls exchanges between applications and services, whatever their purpose. An AI gateway applies functions specific to model requests: model selection, token tracking, costs and usage policies. These are different scopes that can coexist. An AI gateway does not replace the management of your business APIs.

Is Witkeep an AI gateway?

Witkeep includes an AI gateway and combines it with your business tools, contexts, skills and team sharing. This gives you an AI operational center to create deliverables, take action in your tools and track usage, costs and ROI.

We already have an API gateway. Can we use Witkeep?

Yes. Witkeep can fit into an architecture that already has an API gateway. You keep your infrastructure; Witkeep adds AI usage management and a business workspace. The integration covers the AI requests you route through Witkeep.

Our gateway already has an AI module. What does Witkeep add?

Witkeep adds the business layer: work contexts, reusable methods, shared deliverables and conversations, plus usage and value tracking. The aim is to connect technical AI management with the work your teams actually do.

Can we use Witkeep without another gateway?

Yes. You do not need an existing gateway to use Witkeep as an access and control point for your AI usage. Witkeep combines this role with a collaborative workspace. Its scope is AI usage, rather than every API in your information system.

Do we have to use Witkeep’s chat interface?

No. An existing interface can connect to Witkeep if it supports configuring a compatible AI service. Witkeep’s interface brings business contexts, deliverables and team sharing together; these features are not automatically available in a third-party interface.

How can we monitor AI usage in our observability tools?

The dashboard tracks usage, tokens, costs and quotas. Prometheus and OpenTelemetry connect this observability to your monitoring tools.

Let’s see how Witkeep fits your organisation

Request a demo ↗