Turn regressions into fixes
Connect
SPIRITT
Arize AXArize AX
Make evaluations drive better releases
Turn Arize AX traces, datasets, experiments, and prompt versions into a repeatable quality operation for your AI product. SPIRITT brings the spaces, traces, datasets, experiments, and prompts into one working operation, coordinates the next steps across your connected tools, and keeps the outcome in view.
What you can do with Arize AX.
Production evidence becomes test cases
Read bounded traces and spans, isolate concrete failure examples, and add them to an Arize dataset with the context needed to reproduce the problem.
Experiments keep their evidence
Create experiments and record measured runs, then retrieve the run outputs for comparison. SPIRITT coordinates test execution and keeps results tied to the chosen dataset.
Prompt changes stay traceable
Create prompts and immutable prompt versions, inspect their history, and connect each candidate to the experiment evidence used in a release decision.
Works with your other apps.
Get started in three steps.
01
Open your workspace
Create a SPIRITT workspace for your Arize AX operation. Describe the outcome, the records involved, and what finished work should look like.
02
Connect Arize AX
Connect Arize AX with access to the spaces, traces, datasets, experiments, and prompts needed for the mission. Connect companion apps for the handoffs you want SPIRITT to run.
03
Delegate a mission
Ask SPIRITT to turn Arize AX traces, datasets, experiments, and prompt versions into a repeatable quality operation for your AI product. Set the approval points and cadence, then let it carry the work through.
Hand over a mission.
Choose a mission. Make it yours.
Put Arize AX to work.
Make evaluations drive better releases
Questions
Arize AX, with SPIRITT.
01Can production failures become evaluation examples?+
Read bounded traces and spans, isolate concrete failure examples, and add them to an Arize dataset with the context needed to reproduce the problem.
02Can it run an evaluation program?+
Create experiments and record measured runs, then retrieve the run outputs for comparison. SPIRITT coordinates test execution and keeps results tied to the chosen dataset.
03How are prompt revisions tracked?+
Create prompts and immutable prompt versions, inspect their history, and connect each candidate to the experiment evidence used in a release decision.
04Does creating an experiment execute the model?+
Creating an Arize AX experiment stores its runs and associations. SPIRITT coordinates model execution in your connected application or test harness and records the actual outputs. A stored experiment or prompt version is not evidence of a completed model run.
05How do I set up Arize AX with SPIRITT?+
Open a SPIRITT workspace, connect Arize AX, and describe the mission. Choose the spaces, traces, datasets, experiments, and prompts it should use and the actions it can take. Add the companion connections for messages, records, or deliverables outside Arize AX.
06How does SPIRITT follow through on Arize AX work?+
A release review uses actual experiment outputs, the selected dataset, and the recorded prompt version. SPIRITT links failing examples to engineering work and checks later evaluation evidence before closing the regression.