Built for agent runtimes

AI Agents Hub

OpenAI-compatible · the same models (Kimi-K2.6 / Qwen3 / MiniMax) cheaper · pay by card or crypto, no KYC · no subscription.

Pick your tool below for a step-by-step setup page, or jump to the universal config generator.

Whatever your app is — it only needs these 3 values.

Open its settings, find these fields (names vary a little), paste, save. That’s the whole setup.

Base URL (API URL, Endpoint) https://router.mingles.ai/v1
API key (Token, Secret) Create yours here — takes 30 seconds, no card
Model (Model ID) moonshotai/Kimi-K2.6 or MiniMaxAI/MiniMax-M2.7

Asked for a display name or provider name? Type anything, e.g. “Mingles Router”. Still didn’t work? Pick your tool below for click-by-click steps, or use the config generator.

Autonomous agents

Personal agents that plan and act on their own.

Hermes

Verified

NousResearch’s personal agent — native custom-endpoint wizard.

Set up Hermes →

OpenClaw

Verified

Open-source personal agent with native custom-provider config.

Set up OpenClaw →

Coding agents / IDE

Editors and CLI agents that write and run code.

OpenHands

Verified

Autonomous coding agent (OpenAI-Compatible via LiteLLM).

Set up OpenHands →

Kilo Code

Verified

VS Code / JetBrains coding agent — OpenAI-Compatible provider.

Set up Kilo Code →

Cline

OpenAI-compatible

VS Code agent — OpenAI-Compatible provider.

Generate Cline config →

Cursor

OpenAI-compatible

Override the OpenAI base URL + custom model.

Generate Cursor config →

Aider

OpenAI-compatible

CLI pair-programmer via OPENAI_API_BASE.

Generate Aider config →

Continue

OpenAI-compatible

VS Code / JetBrains autocomplete + chat.

Generate Continue config →

Workflows

Automation platforms that call models as a step.

n8n

OpenAI-compatible

Workflow automation — OpenAI credential base-URL override.

Generate n8n config →

Universal config generator

Pick your tool. Paste your key. Copy the result into your runtime. Everything runs client-side — nothing is sent anywhere.

Stored only in this browser. If empty, we use the placeholder <your-key>.

config.json

Frequently asked

Which agent runtimes are supported? +

Hermes, OpenClaw, OpenHands, Kilo Code, Cline, Continue.dev, Cursor, Aider, n8n and anything else that speaks the OpenAI Chat Completions API.

Do I need a special SDK? +

No. Use the official OpenAI SDK (Python, Node, Go, anything). Change only the base URL to https://router.mingles.ai/v1.

Which models support tool-calling? +

Both moonshotai/Kimi-K2.6 and Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 support tool-calling. The gateway parses Hermes XML and re-emits OpenAI tool_calls.

Is there a request ID for tracing? +

Yes — every response includes X-Request-Id and X-Devshard-ID. You can also send your own X-Request-Id and we will propagate it.

How do I pay? +

Top up your balance with a card or crypto in the console — no KYC, no subscription. Billing is per token.

Get API key Read the tools guide