Skip to main content
The pages before this one are what you can call today. Three items are not available yet, in this order.

Experiments

Run a published pack against a dataset. The API is live. The screen is the current build.

judged AI

An assistant in the app that returns the call for your system.

Threshold replay

Apply a new review rule to answers already stored.

Order

1

Experiments

Upload tickets, run a published pack, and read accept, review, and the review rate before that version is the live one.
2

judged AI

Ask how to connect your system. The assistant answers from these docs and from the packs on your account.
3

Threshold replay

Change a review rule and see whether stored answers would move. The model is not called again.

Experiments

Datasets and simulations already run over the API. The Experiments item in the app stays disabled until the screen opens.
Create datasets and start runs from the API. The route is documented on Experiments.

judged AI

An assistant in the app. It reads these docs and the packs on your account, then returns the REST, MCP, or webhook call that connects your system to a published pack. There is no assistant in the app today. REST, MCP, and webhooks are the way in. See Integrations.

Threshold replay

A review rule lives on the pack. Today, changing it means a new judgment. Replay applies the new rule to answers you already stored.