How pref0 learns tool and technology preferences

Developers and teams have strong opinions about tools — package managers, deployment targets, databases, CI systems. pref0 learns these from corrections so agents suggest the right tools every time.

What pref0 extracts

Examples of tool preferences that pref0 detects and scores from conversations.

"Use pnpm, not npm"

package_manager: pnpm0.70
Explicit correction

"Deploy to Vercel, not Netlify"

deploy_target: vercel0.70
Explicit correction

"Use Postgres, not MySQL"

database: postgres0.70
Explicit correction

"GitHub Actions for CI"

ci_cd: github_actions0.40
Explicit instruction

"Terraform, not CloudFormation"

iac_tool: terraform0.70
Explicit correction

"Use Docker for everything"

containerization: docker0.40
Explicit instruction

Confidence progression

See how a single preference compounds in confidence across multiple conversations.

Session 1

"Use pnpm instead of npm"

0.70
Session 3

"pnpm install, not npm install"

0.85
Session 5

Agent uses pnpm without being asked

0.95
Session 6

User confirms pnpm usage

1.00

Related preference types

Code Style PreferencesFormatting PreferencesWorkflow Preferences

See preference learning in action

Your users are already teaching your agent what they want. pref0 makes sure the lesson sticks.

Get Started FreeView Docs