> ## Documentation Index
> Fetch the complete documentation index at: https://fairforge.alquimia.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Fair Forge

> Performance-measurement library for evaluating AI models and assistants

<div id="landing-page" className="landing-hero">
  <div className="landing-badge">by Alquimia AI</div>

  <img src="https://mintcdn.com/alquimiaai/obSyWJO9C9y2zyXt/assets/logo-light.svg?fit=max&auto=format&n=obSyWJO9C9y2zyXt&q=85&s=f2b48a59e1254c1a071465cf14de046c" alt="Fair Forge" className="landing-logo light-only" width="480" data-path="assets/logo-light.svg" />

  <img src="https://mintcdn.com/alquimiaai/obSyWJO9C9y2zyXt/assets/logo-dark.svg?fit=max&auto=format&n=obSyWJO9C9y2zyXt&q=85&s=ce342e8c18b43a396d88cd8e879050d3" alt="Fair Forge" className="landing-logo dark-only" width="480" data-path="assets/logo-dark.svg" />

  <p className="landing-subtitle">
    A comprehensive performance-measurement library for evaluating fairness, toxicity, bias, and conversational quality of AI systems.
  </p>

  <div className="landing-buttons">
    <a className="landing-btn-primary" href="/quickstart">Get Started →</a>
    <a className="landing-btn-secondary" href="/introduction">Documentation</a>
  </div>
</div>

<div className="landing-features">
  <div className="landing-features-grid">
    <div className="landing-feature">
      <div className="landing-feature-title">Toxicity Analysis</div>
      <div className="landing-feature-desc">Detect toxic language patterns with demographic profiling and group fairness scoring.</div>
    </div>

    <div className="landing-feature">
      <div className="landing-feature-title">Bias Detection</div>
      <div className="landing-feature-desc">Measure bias across protected attributes using IBM Granite and LlamaGuard guardians.</div>
    </div>

    <div className="landing-feature">
      <div className="landing-feature-title">Conversational Quality</div>
      <div className="landing-feature-desc">Evaluate dialogue using Grice's Maxims for relevance, clarity, and truthfulness.</div>
    </div>

    <div className="landing-feature">
      <div className="landing-feature-title">Test Generation</div>
      <div className="landing-feature-desc">Generate synthetic test datasets from your documentation with pluggable strategies.</div>
    </div>

    <div className="landing-feature">
      <div className="landing-feature-title">Model Comparison</div>
      <div className="landing-feature-desc">Run tournament-style evaluations between multiple assistants with LLM-as-judge.</div>
    </div>

    <div className="landing-feature">
      <div className="landing-feature-title">Explainability</div>
      <div className="landing-feature-desc">Understand model decisions with token-level attribution and attention analysis.</div>
    </div>
  </div>
</div>

<div className="landing-code">
  <h2 className="landing-section-title">Simple by design</h2>
  <p className="landing-section-desc">Evaluate any AI system in a few lines of code.</p>
</div>

<div className="landing-code-wrapper">
  ```python theme={null}
  from fair_forge.metrics.humanity import Humanity

  metrics = Humanity.run(MyRetriever)
  ```
</div>

<div className="landing-cta">
  <h2 className="landing-cta-title">Ready to evaluate your AI?</h2>
  <p className="landing-cta-desc">Install Fair Forge and start measuring in minutes.</p>
  <a className="landing-btn-primary" href="/installation">Install Fair Forge →</a>
</div>