01
Subscription first
Use the accounts and subscriptions your team already pays for before sending traffic elsewhere.
AI routing infrastructure
A control layer between your AI tools, coding agents and model providers.
Routing
BlackRiver Gateway gives teams one endpoint for many providers. Requests can be routed by cost, model capability, quota, latency, availability or internal policy.
When one provider fails, hits limits or becomes too expensive for the task, the gateway can move traffic without forcing every tool to be reconfigured.
01
Use the accounts and subscriptions your team already pays for before sending traffic elsewhere.
02
Move requests to alternative providers when rate limits, outages or quota exhaustion hit.
03
Select models based on task value instead of always defaulting to premium endpoints.
Tools
Coding agents
Route coding traffic through one controlled layer.
AI workflows
Avoid brittle one-provider workflows.
Providers
Keep optionality between major model ecosystems.
Deployment
Run where your security and operational model requires.
Architecture
The gateway exposes an OpenAI-compatible API surface while routing traffic to multiple provider backends behind the scenes.
const client = new OpenAI({
baseURL: "https://gateway.blackriver.ai/v1",
apiKey: process.env.BLACKRIVER_GATEWAY_KEY,
});
const response = await client.chat.completions.create({
model: "auto:code",
messages: [{ role: "user", content: "Refactor this module" }],
});
// Gateway decides:
// provider, model, fallback, quota, cost, policy.
If this product is based on a fork of an open-source project, preserve the original license, attribution and notices in the repository and documentation.
Enterprise
Define allowed providers, model tiers and fallback behavior by team, tool or project.
See which tools, agents and workflows are consuming AI budget through the gateway.
Gateway traffic can feed cost intelligence, audit reports and optimization dashboards in TokenOptimizer.
Contact
Gateway deployment, provider routing, enterprise controls, private forks and integration with TokenOptimizer.