01→
Agent sends a request
Your app or agent calls the OpenAI-compatible endpoint and includes the model plus any payment context it already has.
One OpenAI-compatible endpoint for GPT, Claude, Gemini, Llama, Qwen, DeepSeek, and more. Fund with Polygon USDT today, with more chains and assets coming next, or let agents pay per-request via MPP. No credit card. No bank account.
$0.50 free for every new workspace. Fund with crypto when you need more volume.
Point your existing SDK at EshuRouter, keep the request shape you already know, and fund usage with crypto. That is the shortest path from prototype to production for both apps and agents.
from openai import OpenAI
client = OpenAI(
base_url="https://www.eshurouter.net/api/v1",
api_key="ek-your-key",
)
response = client.chat.completions.create(
model="auto",
messages=[{"role": "user", "content": "Hello"}],
)The product is not just a model catalog. It combines funding, routing, agent payments, and operational visibility into one gateway you can actually ship with.
Top up your workspace with crypto and use one balance across every live model in EshuRouter. Polygon USDT is live today, with more assets coming next.
See funding flowLet agents attach a machine payment to each request so they can pay for inference without a pre-funded dashboard account or human approval loop.
View MPP docsSet `model: "auto"` and EshuRouter picks the lowest-cost live model that fits the request, while preserving the option to pin a specific model whenever you want.
View routing docsTrack credits, per-request cost, model usage, logs, and guardrails in one dashboard so you always know where each dollar of AI spend went.
Open dashboardInstead of forcing every request through a pre-funded dashboard balance, MPP lets an agent receive a payment challenge, attach proof, and continue the same workflow with a verifiable machine-to-machine payment step.
Your app or agent calls the OpenAI-compatible endpoint and includes the model plus any payment context it already has.
If the request should be paid with MPP, EshuRouter returns a machine-readable payment challenge instead of a normal completion.
The agent signs the challenge, attaches the payment artifact, and retries the request without waiting for a human approval loop.
The gateway verifies the payment proof, serves the model call, and records usage so spend stays traceable in one place.
Everything below is designed to reduce friction: fast signup, crypto funding, and an API path that feels familiar from the first request.
Use Google, GitHub, or a Web3 wallet to create a workspace fast. No credit card. No long KYC flow.
Start with the $0.50 signup bonus, then fund with Polygon USDT today. More chains and assets are coming, and agents can use MPP for per-request payments.
Create an API key, keep your existing OpenAI SDK, and point base_url to EshuRouter. The quickstart docs walk you through the first request.
Use leading models through one crypto-funded API, or let EshuRouter auto-pick the best live option.
EshuRouter and MPP are designed for agent-native spending, where software can fund and route inference without a human in the loop.
Manus
Plan work, spend per request, and complete agent loops in one path.
OpenClaw
Plug open-source orchestration into one routing and billing layer.
MuleRun
Combine crypto funding, routing, and logs inside one runtime layer.
Sign in, grab an API key, and hit the chat completions endpoint with your existing OpenAI SDK. No card, no KYC.
View docs →See how crypto funding, machine payments, and per-request settlement fit together for agent-native inference.
View payment guide →Use the $0.50 signup balance first, then fund with Polygon USDT when you need more throughput.
Create an API keyUse one key per app, agent, or environment so spend and traffic stay easy to trace.
Check your first requestSee which model served the call, how long it took, and what it cost inside the logs tab.