One Gateway.
Every Model.
Real Control.

Unified access to OpenAI, Claude and Gemini today — more providers integrating now. Control routing, cost and reliability from a single platform.

Claude Code · Codex · OpenAI SDK5 minute integration
01 — UNIFIED MODELS

Frontier models. One endpoint.

Drop-in OpenAI-compatible. OpenAI, Anthropic and Google are live; more providers in active onboarding.

OPENAI
OpenAI
ONLINE
  • GPT-5.5FRONTIER
  • GPT-5.5 ProPRO
  • GPT-5.4AFFORDABLE
  • GPT-5.3-CodexCODEX
ANTHROPIC
Anthropic
ONLINE
  • Claude Opus 4.7FRONTIER
  • Claude Sonnet 4.6BALANCED
  • Claude Haiku 4.5FAST
GOOGLE
Google
ONLINE
  • Gemini 3.1 ProFRONTIER
  • Gemini 3 FlashFAST
  • Gemini 2.5 ProSTABLE
02 — CAPABILITIES

Engineered for production load.

Six primitives that turn raw LLM access into a controllable, billable, observable platform.

Unified API surface

A single OpenAI-compatible endpoint fronts every live provider. Swap models with one parameter — no SDK churn, no vendor lock-in.

PROTOCOLOPENAI v1
STREAMINGSSE / WS

Multi-key load balance

Pool unlimited upstream keys per provider. Weighted round-robin, automatic failover, hot rotation — no request lost.

STRATEGYWEIGHTED RR
FAILOVER< 50ms

Token billing & credits

Pre-paid credit ledger with per-model pricing. Issue sub-keys to teams or customers; bill in your unit of choice.

BILLINGPRE-PAID
LEDGERIMMUTABLE

Quota & rate control

Per-key RPM, TPM and daily budget caps. Hard limits, soft alerts, tenant-scoped throttles enforced at the edge.

LIMITSRPM · TPM · $
SCOPEKEY · TENANT

Real-time analytics

Dashboards for cost, latency, error rate and model mix. Drill from tenant to single request in two clicks.

RETENTION90 DAYS
EXPORTCSV · API

Built on New API

Powered by the open-source New API gateway — battle-tested across thousands of deployments. Audit every line, extend with plugins, and self-host the entire stack on your own VPC.

FOUNDATIONNEW API
DEPLOYVPC · CLOUD
03 — ARCHITECTURE

A control plane for every model.

Requests enter once, are authenticated, metered, routed and streamed back — typically under 200ms of overhead.

PROVIDERS
04 — INTEGRATION

One backend. Every coding agent.

Drop-in for Claude Code, Codex CLI, OpenAI SDK — and every other tool that speaks the OpenAI or Anthropic protocol. Built on the open-source New API gateway.

POST /v1/messages
# Claude Code → eronmind
export ANTHROPIC_BASE_URL=https://api.eronmind.com
export ANTHROPIC_AUTH_TOKEN=em-sk-...

claude "Build a Next.js page with Tailwind"

Stop juggling API keys.

Get Started
eronmind

The unified AI gateway for coding agents. Built on the open-source New API.

© 2026 ERONMIND LABS