4 projects
prompt-playoff
Prompt optimization and LLM evaluation for local LLMs: picks the prompting technique your task needs, compiles the prompt it implies, measures it on your model, and searches for a better one.
footnote-mcp
MCP server for source-grounded web research: search, structured-data extraction, evidence verification, and browser automation.
pressf
Lazy goldset annotation: agent does the work, you press F
abscissa
A safety-aware MCP server for Linear
