Ragas

aid: ragas-ai
name: Ragas
description: Ragas is an open-source evaluation toolkit for Large Language Model applications, with particular depth on Retrieval
  Augmented Generation (RAG) and agentic systems. Originally created under the Exploding Gradients organization on GitHub
  and now maintained by Vibrant Labs AI, Ragas is a Python library distributed on PyPI under the Apache 2.0 license. It moves
  teams from informal "vibe checks" to systematic evaluation loops by providing objective LLM-based and traditional metrics,
  automated test dataset generation, experiment tracking, and integrations with the broader LLM ecosystem including LangChain,
  LlamaIndex, OpenAI, Anthropic, and popular observability platforms. Ragas exposes a metrics library covering faithfulness,
  response relevancy, context precision and recall, factual correctness, semantic similarity, agent tool-use accuracy, SQL
  equivalence, Nvidia-defined RAG metrics, and general-purpose rubric scoring. The project ships a CLI (`ragas`) with quickstart
  templates such as `rag_eval`, and is consumed primarily as a `pip install ragas` library rather than as a hosted API service.
  Ragas is widely cited as a default evaluation harness for RAG applications and has grown a substantial community on GitHub
  and Discord.
type: Index
position: Provider
access: 3rd-Party
image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/apis-json-logo.jpg
tags:
- LLM Evaluation
- RAG Evaluation
- Retrieval Augmented Generation
- AI Evaluation
- Open Source
- Python
- Metrics
- Test Data Generation
- Agent Evaluation
- LLM Tooling
url: https://raw.githubusercontent.com/api-evangelist/ragas-ai/refs/heads/main/apis.yml
created: '2026-05-25'
modified: '2026-05-25'
specificationVersion: '0.20'
apis:
- aid: ragas-ai:ragas
  name: Ragas Python Library
  description: The Ragas Python library is the primary surface of the project, installed via `pip install ragas` and imported
    as `ragas`. It exposes evaluation entry points (`ragas.evaluate`), metric classes (Faithfulness, AnswerRelevancy, ContextPrecision,
    ContextRecall, FactualCorrectness, SemanticSimilarity, ToolCallAccuracy, AgentGoalAccuracy, and more), dataset generation
    utilities, and integrations with LangChain and LlamaIndex. The library is not an HTTP API — it is consumed in-process
    by Python evaluation scripts, notebooks, and CI pipelines.
  humanURL: https://docs.ragas.io/
  tags:
  - Python
  - Library
  - Evaluation
  - RAG
  properties:
  - url: https://docs.ragas.io/
    type: Documentation
  - url: https://github.com/explodinggradients/ragas
    type: SourceCode
  - url: https://pypi.org/project/ragas/
    type: SDKs
  - url: https://github.com/explodinggradients/ragas/blob/main/LICENSE
    type: License
common:
- type: DomainSecurity
  url: security/ragas-ai-domain-security.yml
- type: Website
  url: https://www.ragas.io/
- type: Documentation
  url: https://docs.ragas.io/
- type: GettingStarted
  url: https://docs.ragas.io/en/stable/getstarted/
- type: Concepts
  url: https://docs.ragas.io/en/stable/concepts/
- type: Metrics
  url: https://docs.ragas.io/en/stable/concepts/metrics/available_metrics/
- type: HowToGuides
  url: https://docs.ragas.io/en/stable/howtos/
- type: SourceCode
  url: https://github.com/explodinggradients/ragas
- type: GitHubOrganization
  url: https://github.com/explodinggradients
- type: Package
  url: https://pypi.org/project/ragas/
- type: License
  url: https://github.com/explodinggradients/ragas/blob/main/LICENSE
- type: Issues
  url: https://github.com/explodinggradients/ragas/issues
- type: ReleaseNotes
  url: https://github.com/explodinggradients/ragas/releases
- type: Discord
  url: https://discord.gg/5djav8GGNZ
- type: Twitter
  url: https://twitter.com/ragas_io
- type: Company
  url: https://www.vibrantlabs.ai/
- type: Contact
  url: mailto:founders@vibrantlabs.com
- type: Features
  data:
  - name: RAG Evaluation Metrics
    description: Faithfulness, Response Relevancy, Context Precision, Context Recall, Context Entities Recall, and Noise Sensitivity
      for retrieval augmented generation pipelines.
  - name: Agent and Tool-Use Metrics
    description: Topic Adherence, Tool Call Accuracy, Tool Call F1, and Agent Goal Accuracy for evaluating multi-step agentic
      systems.
  - name: Natural Language Comparison
    description: Factual Correctness, Semantic Similarity, BLEU, ROUGE, CHRF, Exact Match, and String Presence metrics for
      output comparison.
  - name: SQL Evaluation
    description: Execution-based Datacompy Score and SQL Query Equivalence metrics for text-to-SQL applications.
  - name: General Purpose Scoring
    description: Aspect Critic, Simple Criteria Scoring, Rubrics-based scoring, and instance-specific rubrics for custom evaluation
      criteria.
  - name: Nvidia Metrics
    description: Answer Accuracy, Context Relevance, and Response Groundedness metrics contributed by Nvidia for RAG quality.
  - name: Test Data Generation
    description: Automated synthesis of diverse test datasets covering single-hop, multi-hop, and abstract query types over
      user knowledge bases.
  - name: Experiments
    description: Experiment-first workflow comparing prompts, models, and configurations across datasets with iterative result
      tracking.
  - name: Custom Metrics
    description: DiscreteMetric and decorator-based APIs for defining LLM-judge and rule-based custom evaluation metrics.
  - name: CLI Quickstart Templates
    description: The `ragas quickstart` command scaffolds evaluation projects including the `rag_eval` template for RAG systems.
- type: UseCases
  data:
  - name: RAG Pipeline Evaluation
    description: Scoring retrieval and generation quality in RAG applications across faithfulness, relevance, and context
      fidelity.
  - name: Agent Evaluation
    description: Measuring tool-call correctness, goal completion, and topic adherence in multi-step LLM agents.
  - name: Regression Testing in CI
    description: Running Ragas metrics in CI pipelines to detect quality regressions across prompt, model, and configuration
      changes.
  - name: Model and Prompt Selection
    description: Comparing candidate models and prompt variants on a fixed dataset using Ragas experiments.
  - name: Synthetic Test Set Generation
    description: Generating diverse evaluation datasets from a knowledge base for systematic LLM testing.
  - name: Text-to-SQL Evaluation
    description: Validating generated SQL against reference queries using execution and structural equivalence metrics.
- type: Integrations
  data:
  - name: LangChain
    description: Native integration for evaluating LangChain chains, retrievers, and agents using Ragas metrics.
  - name: LlamaIndex
    description: Integration for evaluating LlamaIndex RAG pipelines and query engines.
  - name: OpenAI
    description: Default LLM judge backend uses OpenAI models such as GPT-4 class judges.
  - name: Anthropic
    description: Anthropic Claude models supported as LLM judges via the LangChain LLM abstraction.
  - name: Hugging Face
    description: Support for Hugging Face embeddings and models as judges, plus dataset interop via the `datasets` library.
  - name: LangSmith
    description: Result tracking and trace inspection via LangSmith observability.
  - name: Arize Phoenix
    description: Observability integration for tracing Ragas evaluations alongside production LLM traffic.
  - name: Helicone
    description: LLM cost and trace observability for Ragas-driven evaluations.
  - name: Pandas
    description: Datasets and evaluation results are exposed as pandas DataFrames for analysis.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com