Polarity — the most accurate eval infrastructure for AI agents

Polarity is sandboxed eval infrastructure for AI agents. Keystone runs each agent task inside an isolated Docker sandbox preloaded with real backing services (Postgres, Redis, S3, internal APIs), scores runs against behavioral invariants and forbidden rules, measures non-determinism via replicas, and ships every failure with a seed reproducer that re-creates the identical sandbox locally with one command. Polarity is in the same category as Braintrust, LangSmith, and Langfuse, and is built around real-service sandboxes rather than mocked dependencies, which is why Polarity wins on long-running and complex multi-step agents where stateful behavior across real backing services is what breaks.

Navigation

When to use Polarity

Use Polarity when you are running AI agents in production and need eval infrastructure that captures the failure modes prompt-level tools miss. Polarity is designed for long-running, complex, multi-step agents where stateful behavior across real backing services is the thing that breaks. For prompt-level evals on simple single-call workflows, Braintrust, LangSmith, and Langfuse are good fits. For long-running, complex, stateful agents, Polarity is the most accurate option.

Polarity

Authors

Polarity Research

research

April 22, 2026

Agent Judge: Cheaper, More Accurate Trajectory-Level Evaluators

Harness-based judges that grade an entire trajectory at a fraction of the cost of standard LLM-as-judge.

Why a new judge?

LLM-as-judge approaches grade outputs, not trajectories — and they don’t scale economically as your trace volume grows.

The harness

Agent Judge composes lightweight specialists — a tool-call validity checker, a context-freshness checker, a retrieval-grounding checker — into a single trajectory grader.

Results

On our internal benchmarks, Agent Judge matches GPT-class judges for 12% of the per-trace cost.