  ___ _    _ __  __ ___ ___
 | _ \ |  | |  \/  / _ \ _ \
 |  _/ |__| | |\/| | (_) |  /
 |_| |____|_|_|  |_|\___/_|_\

versioned prompts. replayable evals.

A prompt registry that diffs like git, an eval bench that runs in CI, and OpenTelemetry traces of every model call — without locking your production traffic into our proxy.

$ signup $ docs

// registry

Prompt versions are content-addressed; rollbacks are an SHA away.

// evals

Replay a frozen test set against a candidate prompt, with diff-of-outputs.

// traces

OTel-native spans for every call; ship them to your existing backend.