TruLens
Open-source LLM evaluation and observability.
About
TruLens (truera/trulens, MIT, ~3.6k stars) is TruEra's open-source framework for evaluating and tracking LLM apps: feedback functions (groundedness, relevance, toxicity), tracing, and dashboards.
Actively maintained (18 releases/12mo). Smaller than Langfuse in adoption, but fully open source and eval-focused — a good fit for teams that want RAG triad metrics without a hosted platform.
What it can do
- Feedback functions (groundedness, relevance)
- Tracing and observability
- Evaluation dashboards
- RAG triad metrics
- MIT open source
Setup
pip install trulens.- Instrument your LLM app with the TruLens recorder.
- Define feedback functions (e.g. groundedness).
- View traces and scores in the dashboard.
Prerequisites
Python 3.9+; an LLM app to evaluate.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier). MIT, actively maintained with a clean record. Docked for smaller adoption (~3.6k stars) than the eval incumbents. Verified tier on health and docs.
★ 3,598 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.8/10
- Maintenance · recency, releases, issue load
- 8.9/10
- Openness · license + docs
- 9.4/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · 3,598 GitHub stars (log-scaled vs catalog peers)
- · last push 1d ago; 18 releases/12mo; 101 open issues
- · manual review: TruEra's open-source LLM eval/observability; active; smaller than Langfuse; license: MIT
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.