LocalMind is a local-first prompt runner designed to automate running prompts and scripts against source files and directories. It provides a consistent CLI interface for running prompt files, inline prompts, and directory-wide processing with outputs logged to timestamped files.

• Run prompt files or inline prompts against single files or directories
• Recursive directory processing with file type filtering
• Dry-run mode for safe testing
• Timestamped output logging
• Configurable defaults via ~/.localmind/config.json

git clone https://github.com/jeffsebring/localmind.git
cd localmind
pip install -r requirements.txt

# Run a prompt file on a file
lcm file ~/.localmind/prompts/refactor.md ~/localmind_test/hello.py

# Run inline text
lcm text "uppercase this file" ~/localmind_test/hello.py

# Run a prompt file on a directory
lcm dir ~/.localmind/prompts/uppercase.md ~/localmind_test/ --ext .py

# Dry-run example
lcm text "uppercase this file" ~/localmind_test/hello.py --dry-run 

# Model inline example
lcm text "uppercase this file" ~/localmind_test/hello.py --model deepseek-coder-v2 

LocalMind looks for configuration in ~/.localmind/config.json. Example:

{
  "default_model": "deepseek-coder-v2:latest",
  "outputs_dir": "~/.localmind/outputs",
  "prompts_dir": "~/.localmind/prompts"
}

LocalMind is designed to be local-first, extensible, and modular. It encourages reproducible automation workflows while giving developers complete control over prompts, source files, and output management.