n8n-nodes-hiway2llm

HiWay2LLM is a smart LLM router that lets you use Claude, GPT-4o, Gemini, Mistral and more — with your own API keys — at cost, with no margin markup.

Save 70–90% on LLM costs compared to bundled AI providers.

What is HiWay2LLM?

HiWay2LLM is a BYOK (Bring Your Own Key) LLM gateway that routes your requests to the best model based on your choice, while handling:

• Unified API — one endpoint, all providers (OpenAI, Anthropic, Google, Mistral, Perplexity)
• Transparent pricing — you pay provider cost + a small platform fee, never a hidden markup
• Smart routing — auto-selects optimal model tier based on your budget and task
• Detailed usage tracking — per-request cost, tokens, model used

Installation

In your n8n instance, go to Settings > Community Nodes and install:

n8n-nodes-hiway2llm

Setup

1. Create a free account at app.hiway2llm.com
2. Generate an API key in Settings > API Keys (format: hw_live_...)
3. In n8n, add a new HiWay2LLM API credential with your key
4. Drop the HiWay2LLM node into any workflow

Operations

Chat

Send a message to any supported model and get a text response.

Parameters: | Field | Description | |-------|-------------| | Model | Choose from 15+ models across 5 tiers | | System Prompt | Optional instruction for the model | | User Message | The prompt to send | | Additional Messages | Inject conversation history (JSON array) | | Temperature | Creativity control (0–2) | | Max Tokens | Response length limit | | Response Format | Text or JSON Object | | Return Full Response | Get the full OpenAI-format object |

Output (default):

{
  "text": "Your answer here...",
  "finish_reason": "stop",
  "model": "anthropic/claude-sonnet-4-6",
  "usage": {
    "prompt_tokens": 45,
    "completion_tokens": 120,
    "total_tokens": 165
  }
}

List Models

Returns all models available on your HiWay2LLM plan, with tier and provider info.

{
  "id": "anthropic/claude-sonnet-4-6",
  "object": "model",
  "owned_by": "anthropic",
  "hiway_tier": "standard",
  "hiway_provider": "anthropic"
}

Supported Models

| Tier | Models | |------|--------| | Ultra-Light | Gemini 2.5 Flash Lite, GPT-4o Mini, Mistral Small | | Light | Claude Haiku 4.5 | | Standard | Claude Sonnet 4.6, GPT-4o, Gemini 2.5 Flash, Mistral Large | | Heavy | Claude Opus 4.7, GPT-5, Gemini 2.5 Pro | | Web Search | Perplexity Sonar, Sonar Pro, Sonar Reasoning |

Example Workflows

Content generation pipeline

HTTP Trigger → HiWay2LLM (Chat, Sonnet) → Airtable (Save)

Cost-aware classification

Webhook → HiWay2LLM (Chat, GPT-4o Mini) → Switch → ...

Research agent

Schedule → HiWay2LLM (Chat, Perplexity Sonar Pro) → Notion

Links

• Dashboard & API keys: app.hiway2llm.com
• Documentation: app.hiway2llm.com/docs
• GitHub: github.com/Black-Mytm/n8n-nodes-hiway2llm
• Issues: github.com/Black-Mytm/n8n-nodes-hiway2llm/issues

License

MIT — © 2025 Mytm-Group