Alayra Nexus is the AI gateway for developers, teams and enterprises. Bring your own provider keys, point every app at one OpenAI-compatible endpoint, and let intelligent routing choose the right model — at exactly the price the provider charges.
nexus · live routing · simulated traffic
click a provider to take it down
waiting for traffic…
Presets for the providers you already pay · speaks OpenAI and Anthropic APIs
Provider presets: Anthropic, OpenAI, Google Gemini, Groq, Mistral, OpenRouter, Hugging Face, Cloudflare Workers AI, Cerebras, Any OpenAI-compatible endpoint. Works with: Claude Code, Cursor, Cline, Roo Code, Continue.dev, OpenAI SDK, Anthropic SDK, curl.
Zero markup
Many AI platforms resell tokens and add a margin you never see itemised. Nexus never touches your token bill: requests go out on your own provider keys, so you pay each provider directly, at their published price.
Pay providers directly. Your contracts, your credits, your committed-use discounts — all preserved.
Every request costed. Per-request USD cost, attributed to the team that made it.
Budgets that bite. Daily, weekly or monthly caps enforced before a request reaches a provider.
What a token markup costs you
Kept in your budget every year
$15,000Illustrative. Set the markup to whatever your current platform charges.
Drop-in
Nexus speaks the OpenAI API and the Anthropic Messages API. Point your SDK, your agents or your coding tools at it — streaming, tools and all — and every request inherits pooling, failover, budgets and analytics.
from openai import OpenAIclient = OpenAI(base_url="https://<your-nexus>/v1", api_key="<team-key>")resp = client.chat.completions.create( model="alayra-nexus-1", # auto-route, or name any model you've configured messages=[{"role": "user", "content": "Hello"}],)print(resp.choices[0].message.content)alayra-nexus-1 lets Nexus choose the model — or name any model you have configured.
Inside every request
01
Every request is authenticated with a scoped team key, then checked against per-key RPM and TPM limits and the team’s budget — atomically, before anything reaches a provider.
02
Nexus walks your tiers — Premium, Standard, Fast — tries the team’s own keys first, keeps multi-turn conversations where the prompt cache is warm, and breaks ties toward the cheapest healthy provider.
03
A per-key circuit breaker benches a failing key with an escalating cooldown — 10s, 20s, 40s, up to ten minutes — and probes it back with a single request. The next healthy option fires instantly.
04
Each request is costed in US dollars and attributed to its team. Analytics, CSV export, Prometheus metrics, OpenTelemetry traces and an append-only audit trail mean finance and engineering read the same numbers.
01 Admit
Built for teams and enterprises
SSO, two-factor auth, team budgets and an append-only audit trail are part of Alayra Nexus under Apache-2.0 — not hidden behind a sales call.
Teams register their own provider keys, used only for their traffic — with optional fall-back to the shared pool, or hard isolation that never touches a credential they didn’t bring.
Daily, weekly or monthly USD caps per team, enforced on the admission path. Over budget means a clean 429 with the spend, the cap and when it resets.
Per-key circuit breakers with escalating cooldown and a single half-open probe. Rate limits are handled separately, as back-pressure — not outages.
Provider keys encrypted at rest with AES-256-GCM. TOTP two-factor, SSO, login lockout, constant-time secret checks and SSRF protection on every outbound call.
A live cost and usage dashboard, CSV export, a Prometheus /metrics endpoint and optional OpenTelemetry tracing — plus a playground that compares routing strategies side by side.
One command on a laptop with SQLite, or Postgres and Redis in production. Encrypted backups restore onto a different gateway — even across database engines.
Two ways to run it
Alayra Nexus is open source and self-hosted. Nexus Cloud is the same gateway, run for you. Either way, requests go out on your keys.
Open source
In developmentSelf-hosted AI gateway. Start on a laptop in one command, then move to Postgres and Redis without losing a key.
Hosted
In developmentOne secure path from your application to every model — without running the gateway yourself.
The Kinetic family
Nexus is the first layer of the Alayra stack. Above it, Kinetic brings the same conviction to the place software is written: local-first, human-approved, auditable.
AI-native IDE
Your codebase indexed on your own machine. An agent that writes a plan you approve, asks before it runs a command, and records everything it did.
Agentic Integrated Development Environment
A standalone, chat-first workspace for agent-driven development: plan with Blueprint, execute on the Kinetic engine, preview your app in the next panel.
The name and the mark
Alayra
The name ends in Ra — the sun, the symbol of clarity and permanence. In the mark, two strokes meet at a single point and the sun rises inside them: one endpoint, held in the light.
The mark, explained01
Your keys, your code and your data stay inside the boundary you define. Sovereignty is structural — not a setting.
02
We build for the system under load at 3am, not the demo. Every claim on this site points to code you can read.
03
We are building infrastructure the next generation will maintain. Decades, not quarters.
Early access
Every Alayra product is in active development. Join early access for Nexus Cloud, Kinetic IDE and Kinetic AIDE — or self-host Nexus today.