SoftprobePerformance reviews for AI agents

Capture Sessions · Review Steps · Improve what matters

I QA AI agents in production

Connect OpenCode or LangChain with one pasteable prompt, capture Sessions and Steps, review Findings, and optionally route alerts to Slack.

Agent QA

I evaluate AI agents

Author suites, run deterministic and outcome-based evals in CI, compare runs, and gate releases — Promptfoo-compatible imports with kernel-owned manifests and thelake storage.

Agent Evaluation

I'm a developer

Attach the Java agent to your service, record real traffic, and replay it as regression tests with dependencies auto-mocked.

Onboard your app

I run the platform / CI

Deploy the Softprobe server with Helm. After each release to a test environment, have the pipeline replay automatically and let the result decide whether it continues.

Replay after deployment

I need observability

Istio Wasm agent and SESSIFY session context feed a business-flow dashboard — deploy on GKE with the platform guides.

Business observability