Features

The whole platform, on one page.

Airelay is an AI relay and router, an AI firewall and an MCP gateway in one product — with the cost analytics and access controls to manage your AI as a team. Every capability below is available with the same flat 4% pricing.

AI Gateway

One OpenAI-compatible API and one key for 45 providers and 2,376 models, with routing that keeps you answering.

OpenAI-compatible /v1/chat/completions — change base_url, keep your SDK
Bring your own provider keys; your provider bills you directly
Keys encrypted at rest or kept in your own Noviqent Vault space
Flat 4% on your provider cost, metered per request
Routing policies: fallback chain, weighted load balancing, lowest latency, lowest cost, lowest usage, priority tiers, semantic routing
Per-key circuit breaker with backing-off cooldown
Response feedback API to report unhelpful answers
Searchable model catalog with price, context window and residency

Explore AI Gateway in detail →

AI Guardrails & Plugins

12 AI plugins you switch on for your whole organisation — no application changes.

PII redaction incl. National Insurance number, UK postcode, Bank sort code, NHS number, UK phone number — with optional restore
Semantic cache with hit-rate stats and one-click clear
Semantic prompt guard and semantic response guard
Third-party guardrails: AWS Bedrock Guardrails, Azure AI Content Safety, Google Model Armor, Lakera Guard or your own endpoint
RAG injection from documents you upload
Prompt compression for long conversations
Request-, token- and cost-based rate limits and budgets
LLM-as-judge quality scoring
Prompt templates, decorators and keyword prompt guard
Vault-backed secrets in AI plugin configuration

Explore AI Guardrails & Plugins in detail →

Agents & MCP

Connect agents to tools from a directory of 15,063 remote MCP servers, through a gateway you control.

Directory of 15,063 remote MCP servers, searchable by category
One-click connect, or add any remote MCP server by URL
MCP gateway with API-key or OAuth 2.0 (Noviqent SSO) auth
A2A agent proxy with agent-card discovery

Explore Agents & MCP in detail →

Observability & Governance

See what every request cost and how it performed, and control who can change what.

Analytics

Requests, p95 latency, errors, tokens, cost and cache hit rate over time — broken down by model, provider, route or status.

Request log

Drill into individual requests to see which model answered, how long it took, what it cost and whether a plugin blocked or cached it.

Quality scores

LLM-as-judge results alongside your latency and cost numbers, so a cheaper route doesn't quietly become a worse one.

Team roles & invitations

Owner, Admin, Developer, Viewer — invite colleagues by email with the access they need.

Audit log

Every configuration change, who made it and when — across keys, AI plugins, routing and agent settings.

Usage & cost

Spend and token usage by model and provider, so you can see where your AI budget goes.

Security

The essentials, on every account.

Two-factor authentication

Optional TOTP 2FA from your account's Security page, with any authenticator app.

Tenant isolation

Your keys, AI plugin configuration and data are scoped to your organisation and enforced on every request.

UK-hosted

The gateway runs on Noviqent's UK-hosted infrastructure, with per-provider residency labelled in the catalog. More →

Try it without a key

A curated set of models with a genuine no-cost provider tier works the moment you sign up — no provider key, no card — capped at 20 requests per organisation per day.

Google · google/gemini-2.0-flash
Google · google/gemini-1.5-flash
Groq · groq/llama-3.1-8b-instant
Groq · groq/llama-3.3-70b-versatile
Mistral · mistral/mistral-small-latest

Put one gateway in front of all of it

Free to sign up, no card needed to try the free models. Connect your own provider keys whenever you're ready.

Get started free Read the quickstart