TruLens

What is TruLens?

TruLens is an open-source library for evaluating and tracing AI agents and retrieval-augmented generation apps. It instruments your application with OpenTelemetry so every step is recorded as a span with inputs, outputs, latency, tokens and cost, then attaches metric scores from LLM judges that also explain why a score is low. Built-in metrics cover agent tool selection, plan adherence and execution efficiency, the RAG triad of groundedness, context relevance and answer relevance, MCP tool calling and quality, and summarization. You can tune a judge with your own rubric, few-shot examples and score range, and compare app versions on a leaderboard to find the point where quality and cost balance. A Streamlit dashboard shows leaderboards, trends, records and side-by-side comparisons, and results can be logged to a local SQLite database, Postgres or Snowflake. TruLens supports Python 3.10 through 3.14 and integrates with LangChain, LangGraph and LlamaIndex. It is free to install with pip, and no paid plans are listed.

Use cases

• Trace every step of an AI agent run
• Score RAG answers for groundedness
• Compare agent versions on quality and cost
• Evaluate MCP tool calling
• Add blocking guardrails to an app
• Log human feedback next to automated scores

Who it's for

• AI engineers
• ML engineers
• Python developers
• Data scientists

Why use TruLens

• Per-step traces make it clear which retrieval or tool call caused a bad answer.
• Explained scores point to what went wrong instead of giving only a number.
• Version comparison helps you balance quality against cost before shipping.
• OpenTelemetry-native traces fit into existing observability setups, and the library is free.

TruLens features

• OpenTelemetry-native tracing with span-level detail
• LLM-judge metrics that return explanations
• Agent metrics for tool selection and plan adherence
• RAG triad of groundedness, context relevance and answer relevance
• MCP tool calling and tool quality metrics
• Judge customization with rubric, examples and score range
• Streamlit dashboard with leaderboard, trends, records and compare
• Logging to SQLite, Postgres or Snowflake
• Ground truth evaluations and streaming app support
• Instrumentation for LangChain, LangGraph and LlamaIndex

TruLens integrations

• LangChain
• LangGraph
• LlamaIndex
• Postgres
• Snowflake
• Streamlit
• OpenTelemetry
• MCP

How to get started with TruLens

1. Install TruLens with pip in a Python 3.10 or newer environment.
2. Instrument your app with the instrument decorator or a framework wrapper such as TruGraph.
3. Define metrics with LLM judge providers and select the inputs and outputs they should score.
4. Run your app so records, traces and scores are logged to your database.
5. Launch the dashboard with run_dashboard to compare versions and inspect individual traces.

TruLens limitations

• Requires Python 3.10 through 3.14, with Snowflake and Cortex packages limited to 3.13 and the NeMo package to 3.12
• Judges that use hosted models incur the cost of those model calls
• The deprecated trulens-apps-nemo package installs only on Python 3.10 through 3.12

TruLens pricing

TruLens is open source, so you can use and self-host it for free.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

TruLens alternatives

The main TruLens alternatives listed on Debuly are Langfuse, Arize Phoenix, Ragas, Arize AX, LightAgent and Opik. The table compares pricing and platforms; order is A–Z.

1. 

   Arize AX

   Observability and evaluation platform for AI agents and LLM apps, with tracing, issue detection and experiments on a free plan.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Arize AX (official site)
2. 

   Arize Phoenix

   Open-source tool for tracing, evaluating and experimenting on LLM apps and agents, free to self-host or run locally.

   Developer platformsAutomation and agents
   Open sourceVisit Arize Phoenix (official site)
3. 

   Langfuse

   Open-source LLM observability and evaluation platform for tracing agents, managing prompts and scoring outputs.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Langfuse (official site)
4. 

   LightAgent

   Compact open-source Python framework for building agents with tools, memory, MCP, tracing, guardrails and multi-step workflows.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit LightAgent (official site)
5. 

   Opik

   Comet's open-source platform for tracing, evaluating and monitoring LLM applications and agents, self-hosted or in the cloud.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Opik (official site)
6. 

   Ragas

   Open-source Python library for evaluating LLM and RAG applications with metrics, experiments and synthetic test data generation.

   Developer platformsCodingResearch and search
   Open sourceVisit Ragas (official site)

TruLens FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from trulens.org · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.