A Claude Code mod that keeps the one-hour prompt cache warm during breaks and asks before a cold send, with its estimated rewrite cost and a Send anyway button. Warming uses model requests and costs tokens. Requires Claude Code 2.1.287+. Local copy of cache-tax 2.2.1 (karanb192/cache-tax).
live-voice
grvermeulen/claude-mods
Hands-free voice for Claude Code: listens through the microphone, sends each spoken request, and reads the reply aloud. OpenAI hears and speaks by default (OPENAI_API_KEY); /listen engine local uses Whisper and a Windows voice. /listen to start. Windows.
openai-models
grvermeulen/claude-mods
OpenAI models in Claude Code on your ChatGPT plan: Sign in with ChatGPT, /gpt picker lists them in the /model picker, and the one you pick answers the session's turns with Claude Code's core tools. Experimental.
autoroute
grvermeulen/autoroute
Routes each turn and subagent to the Claude model the task needs, with a prompt-cache-aware cost check.