prmpt turns coding-agent attention into crypto

prmpt is a clever experiment in making AI-assisted coding partially self-funding, but its success depends on developers trusting an ad layer that reads agent output.

• –Ads appear after replies or in Claude Code’s status line, minimizing workflow disruption
• –Contextual matching targets the problem currently on screen rather than a behavioral profile
• –The open-source client improves auditability, but the backend handling ad matching remains closed
• –Sending final agent replies for matching creates a meaningful privacy and security tradeoff
• –Impression payouts are fractions of a cent, making this beer money rather than a serious income stream