kimchi

n8n community node for Kimchi (kimchi.dev) — OpenAI-compatible LLM router by CAST AI

Package Information

Released: this month
Downloads: 0 weekly / 0 monthly
Latest Version: 0.0.3
Author: CAST AI

Available Nodes

Documentation generating…

Kimchi Chat Model icon
Kimchi Chat Model
Use Kimchi (kimchi.dev) as a chat language model in AI agents and chains

Documentation

@castai/n8n-nodes-kimchi

An n8n community node for Kimchi — the OpenAI-compatible LLM router by CAST AI.

Kimchi lets you access any LLM (OpenAI, Anthropic, Gemini, self-hosted models, and more) through a single API key, with built-in routing, caching, and cost optimization.

Kimchi node in n8n

Features

  • 🌶️ Use any model available in Kimchi as an LLM in n8n AI Agents and Chains
  • 🔑 Single API key for all LLM providers
  • 📋 Dynamic model list — always up to date from your Kimchi account
  • ⚙️ Full control over temperature, top-p, max tokens, response format, and more
  • 🏠 Self-hosted Kimchi support via configurable base URLs

Compatibility

Version 0.1.1+ works with both latest and older n8n releases. On recent n8n (n8n-workflow
≥ 2.40.x) the node honors the instance's egress/SSRF policy; on older versions it falls back
to unrestricted egress (the same behavior as version 0.1.0).

Installation

From npm (recommended)

In n8n, go to Settings → Community Nodes → Install and enter:

@castai/n8n-nodes-kimchi

In a custom Docker image

RUN mkdir -p /home/node/custom && \
    cd /home/node/custom && \
    npm install --ignore-scripts @castai/n8n-nodes-kimchi

ENV N8N_CUSTOM_EXTENSIONS=/home/node/custom

Setup

  1. Sign up at kimchi.dev and grab your API key
  2. In n8n, go to Credentials → New → Kimchi API
  3. Paste your API key — the base URLs are pre-filled and work out of the box
  4. Add the Kimchi Chat Model node to your workflow and connect it to an AI Agent or Chain

Credentials

Field Description Default
API Key Your Kimchi API key —
Chat Base URL OpenAI-compatible proxy endpoint https://llm.cast.ai/openai/v1
API Base URL Management API for listing models https://api.cast.ai/v1/llm/openai

The base URLs only need to be changed if you are self-hosting Kimchi.

Node Options

Option Description Default
Model Model to use (loaded dynamically from your Kimchi account) —
Temperature Sampling temperature (0–2) 0.7
Maximum Tokens Max tokens to generate —
Top P Nucleus sampling (0–1) 1
Frequency Penalty Penalize repeated tokens (-2–2) 0
Presence Penalty Penalize tokens already in context (-2–2) 0
Response Format Text or JSON mode Text
Timeout Request timeout in ms 360000
Max Retries Number of retries on failure 2

Self-hosted Kimchi

If you run your own Kimchi instance, update both base URLs in the credential to point to your deployment:

  • Chat Base URL: your OpenAI-compatible proxy endpoint
  • API Base URL: your management API endpoint

License

MIT

Discussion