Multi-Model Routing
Route each request by rules you write. A rule can use the model, the team, the budget used or the request complexity. Connect the providers you already use, or any OpenAI-compatible endpoint.
Operate · 02 / 15
One door for every model.
Anar Gateway sits between your organization and every AI provider you use. It routes each request to the right model, enforces who may call what, meters spend against budgets, and writes every request to an audit log.
Capabilities
Route each request by rules you write. A rule can use the model, the team, the budget used or the request complexity. Connect the providers you already use, or any OpenAI-compatible endpoint.
Real-time spend tracking per team, key, and model. Set budgets and rate limits per team and key.
Roles decide who can call models and who can change the configuration. Virtual keys decide which models a team can use.
Spread requests across providers by weight. When a provider fails, the Gateway tries the next one in the fallback list.
How it runs
Your own models, on your own provider contracts, converge on one door. Every request leaves it attributed, priced and written to an audit log. Rules you set in the Gateway can block it first.
Works with
A working session with your team: your data, your policies, your perimeter.