axon labs
open-source tools for when an llm has to be right enough to ship.
-
hallucination detection. four methods, one confidence score.
-
tracing, evals, and alerts for llm apps.
-
pull the context that earns its tokens, drop the rest.
-
semantic chunking for rag. boundaries by meaning.
-
find pii, redact it, put it back later.
-
a wearable that records a day and lets you search it.