AI Router · one OpenAI-compatible endpoint · every request to the right model · up to 90% less spend · 500 welcome credits · no card required
AI Router · one OpenAI-compatible endpoint · every request to the right model · up to 90% less spend · 500 welcome credits · no card required
airouter
— About us

A router written
by the people who use it.

AI Router came out of a concrete problem: paying a frontier model to answer “hello”. Instead of picking the model by hand at every point in the code, we moved the decision inside the endpoint.

2026
first line
3
SDK surfaces
1
endpoint
01
— How it started
The story
Italy· in production since 2026

The first commit is dated 21 April 2026: a router, a credit balance and a dashboard. The problem was already clear back then — in everyday code the model name is written by hand, and it stays that way even when the request does not need it.

We built AI Router the way we wanted it for ourselves: an OpenAI-compatible endpoint that reads the request, estimates its complexity with a model that costs nothing, and sends it to the cheapest model able to answer well. With a fallback when a provider goes down, and with the cost written into the response instead of into a spreadsheet.

Today the same endpoint speaks three languages — OpenAI, Anthropic and MCP — and every request leaves a readable line behind.

— At a glance
  • First line of code 21 Apr 2026
  • Based in Italy
  • Endpoint airouter.it/v1
  • Surfaces OpenAI · Anthropic · MCP
  • Routing levels 3 · 1, 3, 5 credits
  • Revenue model credit subscriptions
02
— What we believe
Five working principles
they are also how we decide when to say no

Strong opinions,
loosely held.

i

The most powerful model is not the right one.

Right means the cheapest one that answers well. Everything else is spend that does not change the result.

ii

An automatic decision has to be explained.

Every response says which model answered, at which level, with how many credits, and whether a fallback fired. No black box.

iii

Your prompts are not ours.

We record metadata — model, tokens, cost, latency. The content of requests and responses never lands in our logs.

iv

Automation must not close doors.

You can always force the level, pin the models with a preset, or call one exact model. The default is a convenience, not a cage.

v

Numbers are measured, not promised.

When we say “up to 90%” it is an estimate on the price gap between models, and we say so. We do not pass it off as an SLA.

03
— Milestones
What has shipped

From one router to three
surfaces in four months.

Apr 2026

The first endpoint

Router, credit balance, Stripe billing and dashboard: the first version already routes across several models.

shipped
Jun 2026

Multi-hub orchestrator

Routing stops depending on a single provider: several inference hubs, with automatic fallback between them.

shipped
Jul 2026

Free plan and levels

The free tier stays on the Simple level; forcing a paid level returns a clear error, not a silent downgrade.

shipped
Jul 2026

Router presets

You pin the model for each level and assign the preset to an API key or to the MCP server. Twenty ready-made presets across nine categories.

shipped
Jul 2026

Signed webhooks

Events on requests and credits, HMAC-SHA256 signature, delivery history and manual re-delivery.

shipped
Jul 2026

Routing delegated to the hub

The admin panel picks the models level by level; billing stops flattening onto a single level.

shipped
Jul 2026

Anthropic surfaces and memory

/v1/messages and /v1/responses arrive alongside /v1/chat/completions, plus the memory store with full-text search.

shipped
Aug 2026

Native SSE streaming everywhere

Streaming becomes native on all three surfaces, sharing one conduit core, with no buffering in between.

shipped
04
— In numbers
What the product is today
updated August 2026
1
endpoint
for every provider
3
routing levels
simple · moderate · complex
0
credits to classify
on every request
3
SDK surfaces
OpenAI · Anthropic · MCP
20
ready-made presets
across 9 use categories
1·3·5
credits per level
fixed price, not per token
500
welcome credits
when you verify your email
4
plans
from 0 to 200,000 credits/month

“We did not want another SDK to learn. We wanted code that was already written to stop overpaying, without changing one more line.”

AR
The AI Router team
airouter.it
line to change
1
endpoint
1
— We build in the open

A router written
by the people who use it.

AI Router started because we did not want to pay frontier-model prices to answer “hello”.

What you get right away
sk-
Your own API key
sk-airouter-… · live immediately
500
Welcome credits
when you verify your email
3
Compatible surfaces
OpenAI · Anthropic · MCP