Which risks could break the thesis, and what would warn of them?
Magma faces the risk that its economics remain undefined until a separate agreement is signed: no data engagement starts from the site alone, each requires a separate signed definitive agreement, and the revenue-share percentage is fixed only in that onboarding agreement and is not published, so terms are priced deal by deal and the site commits the parties to nothing.
Magma asks for introductions to frontier labs and to enterprises that need niche fine-tuning. No data is read until the company that built the agent approves the scope.
2Product Overview
What does the product do for its buyer, and how is it delivered?
Magma captures interaction data that an AI agent already generates, packages it for AI labs, and shares revenue with the agent's builder. The dataset includes traces, corrections, and completed workflows, and the company approves the scope before any data is read.
Its buyers are AI labs and researchers. A 20-minute onboarding call becomes available once the intake is submitted.
Magma
Magma licenses packaged AI agent interaction data to AI labs, sharing revenue.