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

# Health

> The per-workflow home for measuring and improving extraction quality

Every workflow has a **Health** tab — the home for measuring how well it performs and driving that up over time. It brings quality metrics, dataset management, and evaluations into one place, so you can measure quality without leaving the workflow.

<Tip>
  **The short version:** the **Overview** shows how the workflow is doing right now. To measure quality deliberately — and compare workflow versions without the ground shifting underneath you — build a **dataset** of documents with known-correct answers and run **evaluations** against it.
</Tip>

***

## What's in Health

<CardGroup cols={2}>
  <Card title="Overview" icon="gauge">
    Health at a glance: accuracy and confidence, accuracy over time, and your weakest fields.
  </Card>

  <Card title="Datasets" icon="database" href="/guides/health/datasets">
    Manage the workflow's dataset — add documents, edit ground truth, and tag files into sub-datasets.
  </Card>

  <Card title="Evaluations" icon="flask" href="/guides/health/evaluations">
    Run evaluations, browse the run list, and compare accuracy across workflow versions.
  </Card>

  <Card title="Optimizer" icon="wand-magic-sparkles">
    Automatic, iterative improvement of a workflow against its dataset. **Coming soon.**
  </Card>
</CardGroup>

<Info>
  Datasets and evaluations cover **Extract** and **Classify** workflows, and are fully **decoupled from production** — dataset edits and evaluation runs never touch your production data.
</Info>

***

## The Overview

The **Overview** gives you a high-level read on workflow health:

* Average **confidence** and average **accuracy**
* Accuracy and confidence **over time**
* The **weakest fields** — where errors concentrate

New to these two numbers? See [Confidence & Accuracy](/guides/workflows/analytics-quality) for what each one means and how to use it.

***

## The quality loop

Health is built around a simple, repeatable loop:

<Steps>
  <Step title="Build a dataset">
    Add representative documents and confirm their correct values as **ground truth**. See [Building a dataset](/guides/health/datasets).
  </Step>

  <Step title="Evaluate">
    Run a workflow version over the dataset and read its accuracy. See [Running evaluations](/guides/health/evaluations).
  </Step>

  <Step title="Refine">
    Improve fields, instructions, or the pipeline based on where it's failing.
  </Step>

  <Step title="Re-evaluate and compare">
    Run again on the same dataset. Because the dataset is fixed and each evaluation is immutable, the difference between two runs is real signal, not noise.
  </Step>
</Steps>

***

## What's next?

<CardGroup cols={2}>
  <Card title="Building a dataset" icon="database" href="/guides/health/datasets">
    Add documents and author ground truth.
  </Card>

  <Card title="Running evaluations" icon="flask" href="/guides/health/evaluations">
    Score a version and compare across versions.
  </Card>
</CardGroup>