For the complete documentation index, see llms.txt. This page is also available as Markdown.

Recipe: Port a LangSmith eval

Recipe — port a LangSmith eval dataset and judge into Stratix.

Steps

  1. Export LangSmith dataset to JSONL.

  2. Upload to Stratix as private benchmark.

  3. Re-author judge using Stratix judge builder.

  4. (Recommended) GEPA-optimize against your LangSmith labels.

client.benchmarks.upload("./langsmith-export.jsonl", name="ported-from-langsmith")

See also

Last updated

Was this helpful?