fakerybakery

Claude Code Kimi Groq — Development skill for Claude Code

Development community intermediate

A basic proxy to use Kimi K2 on Claude Code through Groq.

How to install Claude Code Kimi Groq

This entry records only its repository, not the path inside it, so there is no exact command to give. Open fakerybakery/openbridge and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Code Kimi Groq does

Use models like Kimi K2, DeepSeek V3.1, and GLM-4.5 in Claude Code.

Alternatives in Development

Kimi CLI Agent — Answers one council question as plain text, with no tool access 664 ★
Openclaw Billing Proxy — Route OpenClaw API requests through Claude Code subscription billing instead of Extra Usage 460 ★
Kimicc — 一步命令 npx kimicc 使用 Kimi K2 运行 Claude Code 192 ★

README

Formerly "claude-code-kimi-groq" - now OpenBridge! Legacy code still available in the `cckg-legacy` branch.

OpenBridge

Use models like Kimi K2, DeepSeek V3.1, and GLM-4.5 in Claude Code.

Open-source API bridge that enables any LLM to work with Anthropic-compatible tools.

Quick Start

Quick start using Kimi K2 on Hugging Face:

pip install openbridge

obcli

# Use with GLM-4.5:
OPENAI_MODEL="zai-org/GLM-4.5:fireworks-ai" obcli

By default, OpenBridge uses Kimi K2 through your Hugging Face account. Hugging Face provides a small free quota every month for all users. Hugging Face Pro subscribers receive $2 in credits each month.

You can use ANY OpenAI-compatible LLM API with support for tool calling! Configure this either through environment variables (`OPENAI_MODEL`, `OPENAI_BASE_URL`, `OPENAI_API_KEY`) or a config file. Run `obcli` to see the configuration options.

Usage

# Install the package
pip install openbridge

# Start the proxy
openbridge --port 8323

# Configure your environment
export ANTHROPIC_API_KEY=NOT_NEEDED # If you are not already logged in to Anthropic
export ANTHROPIC_BASE_URL=http://localhost:8323/

# Run Claude Code
claude

CLI options:

--host - Host to bind to (default: 0.0.0.0)
--port - Port to bind to (default: 8323)

Roadmap

Prompt caching - currently, usage of the proxy is not cost effective as prompt caching is not implemented yet. Contributions welcome!
Route haiku models to a smaller model, currently uses GLM-4.5, very expensive!