#14682·phoenix

️ [harbor] harbor evaluations

Author: axiomofjoyCreated Jul 23, 2026Updated Sep 18, 2026
Labelsroadmapc/evals

Native Harbor support for Phoenix

As a user that has an agent (langgraph, claude code agents, deep agents, whatever) I can run harbor with a Phoenix extension or wrapper that the result of running harbor run my-agent results in an experiment(s) with the instructions and reward functions tied to datasets and evals so I can view my results in Phoenix.

  • #15209
  • #15570
  • #15571
  • #15208
  • #14861
  • #16016
  • #16017

Testing PXI using Harbor

  • #14212
  • #14860
  • #14862
  • #14863
  • #14864
  • #16165