4 projects
muteval
Mutation testing for LLM eval suites. Find out whether your evals would actually catch a regression.
docchat-server
Version-pinned documentation retrieval as a Model Context Protocol server. Gives Claude Code / Cursor / any MCP-aware AI grounded answers from the docs of the exact library version your lockfile pins.
toolpicker
ToolPicker - hybrid lexical + semantic tool selection for LLM agents with many tools.
smolAmem
Multi-tier long-term memory for LLM agents. Install: pip install smolAmem. Import: import mneme.