llmshift runs on your own machine — the client starts it, so there is no endpoint to ping. 1 313 installs a week from pypi. Last commit 25 Oct 2026.
Find retiring LLM calls in code and test the replacement model on real traffic before switching.
Nothing serious here today
Today is the operative word: we check llmshift every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Connect this server
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add llmshift -- uvx llmshift{
"mcpServers": {
"llmshift": {
"command": "uvx",
"args": [
"llmshift"
]
}
}
}[mcp_servers.llmshift]
command = "uvx"
args = ["llmshift"]{
"mcpServers": {
"llmshift": {
"command": "uvx",
"args": [
"llmshift"
]
}
}
}{
"servers": {
"llmshift": {
"type": "stdio",
"command": "uvx",
"args": [
"llmshift"
]
}
}
}Optional settings from the registry entry, left out of the block above:
OPENAI_API_KEY (For tools that call OpenAI models (compare, pick, fix, safety). Not needed for scan or models), ANTHROPIC_API_KEY (For tools that call Claude models. Not needed for scan or models), GEMINI_API_KEY (For tools that call Gemini models. Not needed for scan or models).
Alternatives to llmshift
same job, measured the same wayFind the LLM/AI calls you didn't need — where plain code or a regex does the job. No model.
Make a coding agent prove it: run the repo's real tests, typecheck and lint before it says done.
Finds the LLM model IDs pinned in your code that vendors have already switched off, or have dated to
Bounded readonly SHA256 verification with the official MCP SDK. No code execution or model calls.
Read and post to The Agent Billboard on Solana, with spend limits in code and a reasoning log.
Code changes and unit tests by local MLX models, behind a gate a machine can run.
Static analyzer flagging lookahead bias & data leakage in ML/trading code before backtest.
Local LLM does Claude's bulk work: reads logs and files, researches the web, writes test-gated code.
llmshift — questions
Answers built from our own checks of this server.