Create a nomark-ledger.jsonl with a seed meta entry
Append AI context to your project config (CLAUDE.md, .cursor/rules, or .github/copilot-instructions.md)
Add the ledger to .gitignore
Print quick-start steps

Options

npx nomark-engine                  # full setup
npx nomark-engine --context-only   # print AI context block to stdout
npx nomark-engine --no-ledger      # skip ledger creation
npx nomark-engine --help           # show help

What happens next

# install the engine
npm install @nomark-ai/engine

# import your AI conversation history to build your preference ledger
npx nomark import --platform chatgpt --file export.json
npx nomark import --platform claude --file export.json

# view your resolved preference profile
npx nomark profile

# review and confirm low-confidence signals
npx nomark review

AI context

Running npx nomark-engine appends a context block to your AI config file so your coding assistant knows how to use the engine. If no config file exists, pipe it manually:

npx nomark-engine --context-only >> CLAUDE.md

The context block covers the ledger format, signal types, scoring formula, scope matching, and usage patterns — enough for any AI assistant to integrate the resolver on the first prompt.