Skip to main content
kRouter
All providers

OAuth subscription

kimchi/

Kimchi

Kimchi — connect via OAuth.

kimchi.dev

Quick setup

  1. Open the kRouter dashboard at http://localhost:20128

  2. Go to Providers → click the Kimchi tile

  3. Click Connect and complete the OAuth flow in your browser

  4. Test the connection — kRouter auto-refreshes tokens from here on

Use in your IDE

Point any OpenAI-compatible client at kRouter using your local API key.

ide configbash
# Endpoint
http://localhost:20128/v1

# API key (from Dashboard → API Keys)
sk-krouter-XXXX

# Model
kimchi/<model-id>

How Kimchi billing works with kRouter

Kimchi connects with your existing subscription — kRouter signs in with OAuth and reuses the plan you already pay for, so there is no per-token bill. kRouter authenticates it via oauth. 8 models are routable as kimchi/<model>.

Through kRouter, Kimchi serves chat and code completion and image understanding (OCR / vision-to-text). Requests reach it through the same local OpenAI-compatible endpoint as every other provider, so switching to or from Kimchi is a one-line model change in your client rather than a rewrite.

Kimchi models & pricing

Prices are USD per 1M tokens, cheapest first. You are not billed per token on this provider — the figures show what the same model costs elsewhere, which is what routing here saves you.

ModelRoute asInput /1MOutput /1M
MiniMax-M3kimchi/minimax-m3$0.50$2.00
MiniMax-M2.7kimchi/minimax-m2.7$0.50$2.00
Kimi-K2.7kimchi/kimi-k2.7$1.20$4.80
Kimi-K2.6kimchi/kimi-k2.6$1.20$4.80
Kimi-K2.5kimchi/kimi-k2.5$1.20$4.80
Claude Sonnet 4.6kimchi/claude-sonnet-4-6$3.00$15.00
Claude Opus 4.6kimchi/claude-opus-4-6$5.00$25.00
Nemotron 3 Ultra FP4kimchi/nemotron-3-ultra-fp4——

Kimchi FAQ

Is Kimchi free through kRouter?

Kimchi does not bill per token here — kRouter uses your existing subscription via OAuth, so the only cost is whatever you already pay Kimchi directly. kRouter itself is MIT-licensed and free.

How do I use Kimchi with Claude Code or Cursor?

Connect Kimchi in the kRouter dashboard, then point your tool at kRouter's local endpoint (http://localhost:20128/v1) with a kRouter API key. Any OpenAI-compatible client works, and models are addressed as kimchi/<model>.

Which Kimchi models can kRouter route to?

8 models, including MiniMax-M3, Kimi-K2.7, Kimi-K2.6. Every one is addressed as kimchi/<model> from any client.

What happens when Kimchi is rate limited?

kRouter detects the limit, parks that account until its real reset time, and fails over to your next connected account or provider — so the request still completes instead of erroring.