Magma
Magma enables companies building AI agents to monetize their operational data—traces, corrections, and completed workflows—by packaging it into licensable datasets that AI labs pay for. The platform handles the full workflow from intake through licensing, with companies retaining control over data scope and earning a revenue share. The process starts with a two-minute intake form and a 20-minute onboarding call to map the company's data stack.
- Artificial Intelligence
- Software Only
Funding
Raised to date
$500KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Founder details are not available yet.
Product
Problem
Companies building AI agents generate large volumes of operational data—traces, corrections, and completed workflows—that are valuable for training and improving AI systems. This data is typically discarded, leaving companies without a way to monetize it while AI labs struggle to source high-quality, real-world interaction data.
Solution
Magma enables any company building AI agents to monetize its data in a few clicks. The platform packages an agent's traces, interactions, and mistakes into licensable datasets that AI labs pay for, turning otherwise discarded records into a revenue stream. Magma handles the entire workflow: a two-minute intake form captures company and product details, a 20-minute onboarding call maps the company's stack and data path, and Magma manages packaging and licensing. Companies keep control of the data's scope throughout the process and receive a revenue share as the datasets license.
Target Audience
Primary customers are companies building AI agents that produce operational data—traces, corrections, and workflows—including startups and enterprises using observability platforms like Langfuse or data warehouses like Snowflake.
Features
- Data packaging pipeline that converts agent traces, corrections, and completed workflows into structured, licensable datasets
- Support for common data sources including Langfuse, Supabase/PostgreSQL, OpenTelemetry platforms, S3/GCS/Azure Blob, Snowflake/BigQuery/Databricks, and other databases
- Intake form capturing agent session volume, feedback signals (user ratings, edits, retries, human approval, task success), and data storage locations
- Scope control features allowing companies to determine what data is included in licensed datasets
- Revenue share model where companies earn a share as their data licenses to AI labs
- Compliance handling for regulated data, including privacy policy review and de-identification requirements before licensing