Product / UX / Telemetry & observability

TruLens

Open-source library for evaluation and tracing of LLM apps and AI agents. Its differentiator is 'feedback functions' - programmatic evaluators including the RAG triad (groundedness, context relevance, answer relevance) plus toxicity and others - combined with app instrumentation and tracing, making it more eval-first than gateway/observability-first peers.

MIT. Created by TruEra, which Snowflake acquired in 2024; now maintained by Snowflake (the GitHub org remains 'truera'). A separate hosted Snowflake Cortex AI Observability product exists; the library stays fully open. ~3.4k stars, ~39k PyPI downloads/month.

Openness

5 high confidence

5.0

license: MIT(OSI)
source: public(github.com/truera/trulens)
commercial: Snowflake-Cortex-AI-Observability(separate hosted product)
library: not-feature-gated
core-gated: ungated

MIT library that is fully functional standalone; the Snowflake Cortex hosted product is separate, not a gate on the library.

https://github.com/truera/trulens recorded 2026-06-18
MIT, ~3.4k stars, LLM eval + tracing

Adoption

4 medium confidence

4.0

~38.7k PyPI downloads/month (pypistats) alongside ~3.4k stars.

https://pypistats.org/packages/trulens recorded 2026-06-18
38,729 downloads last month

Capability

4 medium confidence

4.0

Eval-first observability; lighter than Langfuse/Helicone on production gateway and prompt management.

https://www.snowflake.com/en/blog/trulens-open-source-ai/ recorded 2026-06-18
Snowflake acquired TruEra; TruLens stays open source

Unchanged since 2026-07-30 (last edited, not re-checked)