One control point for all your AI
Your teams are already building with AI. The question is whether anyone can see what it costs, control who uses what, and keep it running when a provider has a bad day.
Agent Router Enterprise gives you that control. It sits between the applications, teams, and AI agents that consume AI models and the providers that supply them (OpenAI, Anthropic, Google, and the rest) so every AI request travels through one path you govern, observe, and trust.
It is not a model, and it does not compete with the providers. It is the layer that makes all of them usable on your terms: one interface for the people building with AI, one set of controls for the people accountable for cost, security, and compliance, and one source of truth for what is actually being used.
And it runs in your own environment. Your AI traffic, your data, and your credentials never leave your network. Tetrate hosts only the configuration and analytics that keep the platform easy to run.
What changes when using Agent Router?
Most teams adopt Agent Router Enterprise after AI is already in production, once the early, improvised approach has started to cost them.
Six things that change
Use any provider, switch any time
Every provider is integrated, priced, and maintained differently, and the day one retires a model, you are rewriting code in every app that used it. Tetrate Agent Router hides those differences behind a single endpoint. Add a provider, swap one out, or split traffic across several, all as a configuration change, never a development project. Your choice of provider becomes a business decision again, not an engineering constraint.
Finally see what AI costs
Without a gateway, your AI usage is scattered across provider dashboards, application logs, and spreadsheets. Simple questions, like which team is spending the most this month or which model throws the most errors, turn into investigations. Agent Router Enterprise records every request in one place and shows you the cost and usage as it happens. Pipe the same data into the monitoring tools you already run.
Govern access by policy, not by trust
Left alone, AI access sprawls: developers hold raw provider keys, nobody decides which models are approved for which work, and budgets show up only after the money is spent. Tetrate Agent Router flips that. Manage models, providers, and credentials in one place. Grant each team or application its own access and its own limits. Enforce budgets and rate limits automatically. Capture every administrative action in an audit log. Governance becomes a setting, not an act of faith.
Stay up when a provider goes down
Production AI cannot hinge on a single provider staying healthy. When one slows down or starts returning errors, apps without a backup plan fail or behave unpredictably. Tetrate Agent Router reroutes automatically: define an ordered list of alternatives, and the gateway moves to the next one the moment a provider fails. Your application still gets a clean response. Your operations team sees exactly what happened.
Be ready for AI agents
AI agents increasingly reach tools and data through Model Context Protocol (MCP) servers, one for files, one for tickets, one for documentation, and on it goes. Wire each agent to each server by hand and you get unmanageable configuration and no control over what is reachable. Agent Router Enterprise puts every MCP server behind one endpoint and governs it with the same identity, audit, and access controls as everything else. You get the upside of agents without losing oversight of them.
Mix central and team-owned credentials
Real organisations run on a mix of central and team-owned AI accounts. A central team funds the bulk of usage, while a business unit insists on its own provider account for compliance. Tetrate Agent Router supports Bring Your Own Key (BYOK): a team's own credentials sit alongside the centrally managed ones, and policy decides which to use, not a line of code buried in an application.
How it fits together
Two ideas shape how you run the platform, and both matter to whoever is accountable for it.
Data plane
Runs in your environment
Handles live AI traffic inside your own network. Sensitive data and credentials never leave your infrastructure.
Management plane
Hosted by Tetrate
Holds configuration, policies, audit history, and analytics. Tells the data plane how to behave and collects usage data back. No application content passes between the two.
The second split is between the platform's two applications, the Agent Router Console and the Admin Dashboard.
- The Console is where developers build against the platform.
- The Admin Dashboard is where operators run it.
Same system underneath, two deliberately different views. The Who this documentation is for page shows which application fits which audience.
To get a peek under the hood, check the Architecture Overview, which walks through the data plane, the management plane, and the path an AI request takes between them.
What you get
The platform's capabilities fall into four areas.
| Area | What you get |
|---|---|
| Routing | One endpoint to 200+ models across OpenAI, Anthropic, Google, Azure, Mistral, and more; automatic fallback between providers; weighted traffic splitting; routing rules based on request attributes |
| Identity and access | Per-team and per-application access with its own limits; enterprise single sign-on; role-based access in both the Console and the Admin Dashboard; centrally managed BYOK credentials |
| Observability | Request logs, usage analytics, and cost reporting in the Admin Dashboard; export to the monitoring tools you already run; audit logging for every administrative action |
| Agent infrastructure | MCP aggregation behind a single endpoint; tool catalogues usable by AI clients such as Claude Code, Cursor, and VS Code; a Playground for testing prompts against any configured provider |
That is the short version. The Platform Capabilities page expands each area into the detail that matters when you evaluate the product or plan a rollout.
Where to go next
Continue reading
Key concepts
Defines the vocabulary used across the documentation, including data plane, fallback policy, BYOK, and MCP profile.
Architecture overview
Walks through the data plane, the management plane, and the path a request takes between them.
Get started
Covers prerequisites and installation on AWS, Azure, and GCP.
Guides for developers
Jump straight to the developer workflows if your role is already clear.
Guides for platform operators
Jump straight to the operator workflows if your role is already clear.