Skip to main content

What is TokenVue?

TokenVue is an AI gateway and infrastructure control plane that helps teams securely manage, route, monitor, and optimize LLM traffic across providers and environments.

It sits between applications and AI providers such as OpenAI, Anthropic, Gemini, DeepSeek, and OpenAI-compatible self-hosted models.

Why TokenVue Exists​

Modern AI applications often rely on multiple providers, model families, environments, and routing rules. Without a control layer, teams usually end up spreading provider keys, limits, logs, and failover logic across application code and operations runbooks.

That creates problems around:

  • Reliability
  • Security
  • Observability
  • Governance
  • Cost optimization
  • Vendor lock-in

TokenVue centralizes those controls so teams can operate LLM traffic consistently at production scale.

Core Capabilities​

Virtual Keys​

Virtual Keys are application-facing API keys that route requests through TokenVue instead of exposing provider credentials directly.

Use Virtual Keys to manage:

  • Service ownership
  • Provider and model access
  • Daily, monthly, token, or request limits
  • Guardrail behavior
  • Usage tracking and request monitoring

LLM Config​

LLM Config stores provider-facing setup inside TokenVue.

Use LLM Config to define:

  • Provider name
  • Provider API key
  • Base URL for OpenAI-compatible providers
  • Model name
  • Budget mode
  • Token or request limits

Auto Router​

Auto Router applies routing policy when traffic should move away from a primary provider or model.

Routing decisions can use signals such as:

  • Budget usage
  • Latency
  • Provider health
  • Error rates
  • Quota limits
  • Region or capacity requirements

Guardrails​

Guardrails run before requests are forwarded to providers.

TokenVue supports controls such as:

  • Hard budget caps
  • PII scrubbing
  • Toxicity filtering
  • Prompt injection detection
  • Keyword blocking

Insights, Logs, and Breaches​

TokenVue records gateway activity so teams can review:

  • Token usage
  • Estimated spend
  • Provider latency
  • Model routing behavior
  • Request status
  • Guardrail blocks
  • Failed or suspicious traffic

OpenAI-Compatible API​

TokenVue provides an OpenAI-compatible API layer, so existing OpenAI SDK workflows can use TokenVue with minimal code changes.

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.TOKENVUE_API_KEY,
  baseURL: "https://api.tokenvue.ai/v1",
});

For local development, your gateway base URL may be:

http://localhost:3000/v1

Built for Production AI​

TokenVue is designed for:

  • AI startups
  • SaaS applications
  • Enterprise AI platforms
  • AI agents and copilots
  • Internal AI infrastructure teams
  • Hybrid AI deployments

Whether you are building customer-facing AI products or internal AI systems, TokenVue provides the infrastructure layer needed to operate LLM access securely and reliably.

Next Steps​