Agent Data Synthesis
What it is
A local-first Python framework for generating, executing, and validating agent training data with strict evidence requirements. It processes domain-specific tasks (like Contacts and Workspace) through a pipeline that ensures executable state, tool evidence, and lineage before accepting records.
The problem it solves
Standard synthetic data generation often lacks executable verification, lineage, and quality evidence, leading to unreliable agent training data.
What you could use it for
- Generating high-quality, verifiable training data for LLM agents
- Validating agent tool-calling capabilities against synthetic environments
- Creating reproducible, lineage-tracked datasets for agent evaluation
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.