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

# Overview

> How Galtea's concepts connect

## How Everything Connects

```mermaid theme={"system"}
%%{init: {"theme": "neutral"}}%%
flowchart TD
    P[Product]
    T[Tests]
    Spec[Specifications]
    V[Versions]
    EC[Endpoint Connections]
    TC[Test Cases]
    S[Sessions]
    Tr[Traces]
    IR[Inference Results]
    E[Evaluations]
    M[Metrics]

    P --> |Defined by| Spec
    Spec -.-> |Challenged by| T
    T -.-> |Groups| TC
    TC -.-> |Simulated as| S
    S --> |Are a set of| IR
    IR --> |Hold| Tr
    S --> |Judged in| E
    Spec --> |Evaluated with| M
    E --> |Scored by| M

    P --> |Iterated with| V
    V --> |Linked to| EC
    V --> |Evaluated in| S
```

<Info>The dotted lines represent relationships that don't exist for production-based evaluations, where there is no Test or Test Case and the Session is generated from real user interactions.</Info>

<Note>The diagram focuses on core concepts. Some related concepts are omitted for clarity — see the [Additional Notes](#additional-notes) section.</Note>

## At a Glance

Galtea's concepts split into four lifecycle scopes:

* **What you're testing**: Your [Product](/concepts/product), defined by a set of [Specifications](/concepts/product/specification) and iterated on through new [Versions](/concepts/product/version).
* **What you test with**: [Tests](/concepts/product/test) that group [Test Cases](/concepts/product/test/case); your Product's performance against them is measured with [Metrics](/concepts/metric).
* **What your Product generates**: [Sessions](/concepts/product/version/session), each a sequence of [Inference Results](/concepts/product/version/session/inference-result) along with their [Traces](/concepts/product/version/session/trace).
* **How you measure performance**: [Evaluations](/concepts/product/version/session/evaluation) that apply a [Metric](/concepts/metric) to a [Session](/concepts/product/version/session) or to a specific [Inference Result](/concepts/product/version/session/inference-result).

## Additional Notes

A few relationships are easy to misread, and some are omitted from the diagram for clarity. These notes fill in the gaps:

* **Specifications are the glue.** A [Specification](/concepts/product/specification) links to one or more [Metrics](/concepts/metric) so it can be evaluated, and [Tests](/concepts/product/test) can be generated from a Specification to challenge it.
* **Test Case ↔ Session.** A [Test Case](/concepts/product/test/case) belongs to a [Test](/concepts/product/test). When a [Session](/concepts/product/version/session) is generated to simulate a Test Case, it references that Test Case — so the same Test Case can show up across many Sessions (for example, re-running a Test across multiple Versions). Sessions can also be generated without a Test Case, for example when [recording real user interactions](/sdk/tutorials/monitor-production-responses-to-user-queries) to evaluate the Product in production.
* **Endpoint Connections shape Test generation.** When you create a Test with an [Endpoint Connection](/concepts/product/endpoint-connection) selected, the generated Test Cases are produced to match that endpoint's expected input format.
* **Per-turn evaluations.** An [Evaluation](/concepts/product/version/session/evaluation) can target a single [Inference Result](/concepts/product/version/session/inference-result) instead of the whole Session — useful when you want to score, for example, how the agent's output reflected the retrieval context the RAG system provided on that turn. Per-turn Evaluations still belong to the Session that contains the turn; they additionally point at the specific Inference Result being scored.
* **Two types of model.** The standard [Model](/concepts/model) is what your Product runs on — linked to a Version and used to track costs, token usage, and similar. The **evaluator model** is the LLM Galtea uses to perform the judgment for non-deterministic Metrics; it's selected by name on the Metric, not on the Version. See [Models vs. Evaluator Models](/concepts/model#models-vs-evaluator-models) on the Model page for the full distinction.
* **[User Groups](/concepts/user-group) for human evaluation.** A User Group routes [Human Evaluation](/concepts/metric/evaluation-types#human-evaluation) Metrics to a specific set of annotators, so only the right reviewers see those Evaluations on the dashboard.
* **[Monitors](/concepts/product/monitor) watch production automatically.** A Monitor is an always-on rule on a [Product](/concepts/product) (optionally one [Version](/concepts/product/version)) that scores a sampled fraction of its production [Sessions](/concepts/product/version/session) with a set of metric families, with no run to trigger by hand.

## All Concepts

<CardGroup cols={3}>
  <Card title="Product" icon="box" iconType="solid" href="/concepts/product">
    A functionality or service being evaluated
  </Card>

  <Card title="Specification" icon="bullseye" iconType="solid" href="/concepts/product/specification">
    A testable behavioral expectation for a product
  </Card>

  <Card title="Version" icon="code-branch" iconType="solid" href="/concepts/product/version">
    A specific iteration of a product
  </Card>

  <Card title="Endpoint Connection" icon="plug" iconType="solid" href="/concepts/product/endpoint-connection">
    A reusable connection to an external endpoint for AI product evaluation
  </Card>

  <Card title="Monitor" icon="gauge-high" iconType="solid" href="/concepts/product/monitor">
    An always-on evaluation rule for production sessions
  </Card>

  <Card title="Test" icon="clipboard-list" iconType="solid" href="/concepts/product/test">
    A set of test cases for evaluating product performance
  </Card>

  <Card title="Test Case" icon="list" iconType="solid" href="/concepts/product/test/case">
    Each challenge in a test for evaluating product performance
  </Card>

  <Card title="Session" icon="clock-rotate-left" iconType="solid" href="/concepts/product/version/session">
    A full conversation between a user and an AI system.
  </Card>

  <Card title="Inference Result" icon="arrow-right-from-bracket" iconType="solid" href="/concepts/product/version/session/inference-result">
    A single turn in a conversation between a user and the AI.
  </Card>

  <Card title="Trace" icon="sitemap" iconType="solid" href="/concepts/product/version/session/trace">
    Understand how traces capture the internal operations of your AI agents.
  </Card>

  <Card title="Evaluation" icon="clipboard-check" iconType="solid" href="/concepts/product/version/session/evaluation">
    The assessment of an evaluation using a specific metric's criteria
  </Card>

  <Card title="Metric" icon="gauge" iconType="solid" href="/concepts/metric">
    Ways to evaluate and score product performance
  </Card>

  <Card title="Model" icon="microchip" iconType="solid" href="/concepts/model">
    Way to keep track of your models' costs
  </Card>
</CardGroup>