Open source · Built for development teams

Govern AI coding access. See where usage goes.

Zentrola sits between clients such as Codex and Claude Code and your model providers. Developers keep their tools. Teams manage models, access, credentials, quotas, and costs in one place.

Tools your team already uses Codex / Claude Code / Apps Virtual Keys and App Keys
ZZentrola
Identity & accessToken quotasLogical modelsRouting & records
Your model resources OpenAI / Anthropic / Compatible providers Provider credentials stay server-side

Connecting a model is easy. Governing its use takes more.

Once people, apps, and providers share the same setup, access, credentials, and spending need clear ownership.

Credentials spread everywhere

Real provider API keys get copied into laptops and app settings, making rotation and revocation difficult.

Zentrola issues separate keys to users and apps while keeping upstream credentials centrally.

Model access lacks boundaries

Teams need to know who can use each model and apply the same rules across different clients.

Grant logical models through Groups; enforce permissions and quotas at the gateway.

Costs have no clear owner

A provider total alone does not show which people, apps, or models drove the usage.

Attribute usage to users, apps, models, and resources, with cost details and monthly quotas.

One control plane between existing tools and models.

Keep the AI coding clients your developers know. Zentrola brings the management and model gateway together.

Connect

Serve common OpenAI and Anthropic APIs, with protocol conversion and provider failover.

Govern

Manage providers, credentials, logical models, user and app access, and independent monthly Token quotas.

Account

Inspect usage, operation records, and cost attribution. Filter billing details and export CSV.

Keep administration in one place.

Administrators configure resources and access through a separate Web console. Developers call authorized models from their own clients.

Read the client configuration guide
Sign-in screen for the Zentrola Admin Web console
Admin Web console

Start with one governed model route.

Install Zentrola, configure a provider and logical model, then issue a key to a user or app.