Chat and reasoning
Assistants, agents and long-form thinking with frontier language models.
Unified AI gateway
Every AI model. One archway.
One API key, one endpoint and one token balance for the world's leading AI models. Top up once, switch models in a single line, and stop juggling five provider bills.
1 key for every modelPay as you go0 subscriptions
base_url ="https://ai.archesrouter.my.id/v1"
that is the whole migration
GPT, Claude, Gemini, DeepSeek, Llama, Mistral, Qwen, Grok
Why a router
Five providers means five of everything.
Every model you add brings another account, another key, another invoice and another thing that can go down at 2 a.m. ArchesRouter folds all of it into one doorway.
Without a router
Accountsone per vendor
API keysone per vendor
Invoices / token balanceone per vendor
Request formatall different
When a provider is downso are you
With ArchesRouter
Accounts1
API keys1
Invoices / token balance1
Request format1, OpenAI-style
When a provider is downtraffic reroutes
Models
The right brain for every job.
No single model wins at everything. Use a heavyweight for reasoning, a fast one for chat and a cheap one for bulk work, all from the same key.
Code
Plug your key into coding agents and IDE tools that accept an OpenAI-style endpoint.
Vision
Send screenshots, receipts and documents. Get structured answers back.
Image generation
Product shots, illustrations and social visuals from a prompt.
Embeddings and search
Power RAG, semantic search and recommendations over your own data.
Speech
Transcribe calls and meetings, or give your product a voice.
Integrate
Switch in one line.
Already using the OpenAI SDK? Point it at ArchesRouter and keep every other line of your code. Trying a different model is a string change.
- Drop-in. Works with tools and libraries that accept a custom base URL.
- No lock-in. Change the model ID, not your architecture.
- Minutes, not sprints. Sign up, copy your key, send a request.
from openai import OpenAI
client = OpenAI(
base_url="https://ai.archesrouter.my.id/v1",
api_key="YOUR_ARCHESROUTER_KEY",
)
reply = client.chat.completions.create(
model="openai/gpt-4o",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)
Routing
When one road closes, your request takes another.
Providers have outages and rate limits. Your users should never notice. Break a route below and send a request to watch it recover.
Ready. Press “Send a request”.
Built for builders
Less plumbing. More shipping.
One balance for everything
Buy tokens once and spend them on any model. No more guessing which provider wallet is about to run dry.
A second route, always
When a provider slows down or goes dark, your traffic has somewhere else to go. Your launch day should not depend on someone else's status page.
See every token
Track what you spend per model and per key, so there are no surprises at the end of the month.
Keys you can hand out
Create separate keys for each app, teammate or client and revoke them whenever you need.
No subscription
Pay as you go. Use a lot this week and nothing next week. You only pay for what you send.
Tokens
How far does your balance go?
Tokens are the fuel of every AI request. Drag the slider to see roughly what a top-up turns into.
1Mtokens
750Kwords of text
1.5Kpages (500 words)
3.3Kchat replies (~300 tokens)
Rough estimates for English text. Real usage depends on the model and language.
Get started
From sign-up to first response in three steps.
Create your account
Sign up on the ArchesRouter console. It takes about as long as reading this card.
Top up tokens
Add balance in the amount that fits your project. No plan to pick, no contract to sign.
Call any model
Copy your key, set the base URL and send your first request. Switch models whenever you like.
FAQ
Questions, answered.
Still unsure? Create an account and look around. You only pay when you top up.
What is ArchesRouter?
A unified AI gateway. You get one API key and one token balance, and use them to call models from many AI providers through a single endpoint.
Do I need my own accounts with each AI provider?
No. ArchesRouter handles the provider side. You manage one account, one key and one balance.
Will it work with my existing code?
If your code or tool speaks the OpenAI-style API and lets you set a base URL, change the base URL and the key. Everything else stays the same.
How does billing work?
It is prepaid and pay-as-you-go. Top up a token balance, and each request draws from it based on the model you call. There is no monthly subscription.
How do I start?
Create an account, top up, copy your API key and send a request. Most developers are up and running in a few minutes.
Walk through
Your next model is one line away.
Stop managing providers. Start shipping with all of them.