> ## Documentation Index
> Fetch the complete documentation index at: https://arize-ax.mintlify.dev/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Alyx
> Alyx is the AI agent built into Arize AX. Just as Cursor and Claude Code changed how developers write software - by understanding your codebase and acting on what you ask - Alyx does the same for AI engineering. It already knows your traces, prompts, datasets, and evals, so you can just tell it what you need and it handles it, without mastering every tool, query language, or configuration in the stack.
## Alyx meets you where you are
Alyx adapts to where you are in the platform. Here's where to find it and what it can do:
| Surface | Where to find Alyx | Skills available |
| ------------------------------------------------------------------------- | ------------------------------------- | -------------------------------------------------------------- |
| [**Trace slideover**](/ax/observe/tracing) | Open Alyx from a trace detail view | Trace troubleshooting, span analysis, annotations, build evals |
| [**Prompt Playground**](/ax/prompts/prompt-playground) | Alyx chat in the Playground | Optimize prompts, build evals, run experiments |
| [**Eval Builder / Task Builder**](/ax/evaluate/evaluators) | Alyx on the eval or task builder page | Build custom evals, configure tasks |
| [**Traces table search bar**](/ax/observe/tracing/view-and-manage-traces) | AI Search in the filter bar | Natural language to filter syntax |
| [**Traces page**](/ax/observe/tracing) | Alyx on the main Traces page | Multi-trace analysis, pattern discovery |
| [**Datasets / Experiments**](/ax/develop/datasets-and-experiments) | Alyx on Datasets or Experiments | Analyze experiments, manage datasets |
## Getting Started
### How to open Alyx
Use the keyboard shortcut to open or close Alyx from anywhere in the app.
* **macOS:** **Cmd+L**
* **Windows / Linux:** **Ctrl+L**
When you have text selected, the same shortcut opens Alyx and adds that text as context.
### Adding context to Alyx
Alyx works best when it knows exactly what you're looking at. There are three ways to give it context.
* **Highlight** - Select any text on the page (a span attribute, an error message, a prompt snippet) then press **Cmd+L** on macOS or **Ctrl+L** on Windows/Linux. Alyx opens and adds the selected text to your message. If nothing is selected, the shortcut just opens or closes Alyx.
* **Mention** - Type **`@`** in the Alyx input to open a menu. Mention a dataset, experiment, project, or span and Alyx receives the IDs so it can scope the conversation to that data.
* **Type it** - Include context directly in your message. Reference a trace, dataset, or experiment by ID, or start with "Additional context:" and add what Alyx needs to know.
Whatever you add is sent with your message as input context. Alyx uses it to scope its response to the data you care about so you're not re-explaining your situation every time.
### Alyx Settings
Edit your hotkey from the Settings modal, accessible through the menu in the top right corner of the Alyx chat interface. More settings will be added over time.
## Alyx across your workflow
A single trace tells you something failed. It doesn't tell you why or how widespread it is. Alyx is grounded in live trace and span context including inputs, outputs, tool calls, errors, and your selected time range, so you can go from noticing a problem to understanding its scope in one conversation.
### Where to find Alyx
On the [**Traces page**](/ax/observe/tracing), in the **trace slideover**, and as **AI Search** in the traces table filter bar ([view and manage traces](/ax/observe/tracing/view-and-manage-traces)).
### Skills
| Theme | Skill | Description |
| --------------------------- | ----------------------------- | ------------------------------------------------------------------- |
| **Diagnose & understand** | **Trace preview** | Get an overview of the trace structure and span hierarchy |
| | **Find in trace** | Search for specific content across the trace |
| | **Display query params** | Show the current filter and time range |
| | **Span data** | Inspect detailed input/output, latency, and attributes for any span |
| **Assess quality** | **Annotations** | Create configs and annotate single or multiple spans with labels |
| | **Build eval** | Write a custom LLM-as-a-judge eval for your use case |
| | **Choose evals** | Select or attach evals; list dataset evals and online tasks |
| **Curate training data** | **Create dataset from spans** | Create a dataset from selected trace spans |
| | **Append spans to dataset** | Add spans from this trace to an existing dataset |
| | **Synthetic datasets** | Create or append synthetic data for testing |
| **Version & reuse prompts** | **Prompt Hub** | List, load, and save prompts; save new versions |
| **Find anything fast** | **Search** | Build filters or find spans via natural language |
### Example prompts
* "What's causing the latency in this trace?"
* "Annotate this span as a hallucination"
* "Build an eval to check if the response answers the question"
* "Find all spans that call the search tool"
* "Create a dataset from the spans with errors"
Prompt iteration involves more than writing prompts. It means attaching datasets, mapping columns, wiring evals, running experiments, and keeping track of what changed between runs. Alyx holds the full context of your prompt, dataset, evals, and experiment history in one thread so each iteration builds on the last.
### Where to find Alyx
Alyx is available in the [**Prompt Playground**](/ax/prompts/prompt-playground), and on [**Datasets and Experiments**](/ax/develop/datasets-and-experiments) for comparing runs, inspecting eval breakdowns, and scaling what you learned from traces into repeatable tests.
### Skills
| Theme | Skill | Description |
| ------------------------ | ----------------------------------- | ------------------------------------------------------- |
| **Set up your data** | **List datasets** | List available datasets |
| | **Dataset preview** | Inspect dataset structure and sample rows |
| | **Attach dataset / select columns** | Attach a dataset and choose which columns to use |
| **Curate training data** | **Create synthetic dataset** | Create a new dataset with synthetic examples |
| | **Append synthetic data** | Add synthetic examples to an existing dataset |
| | **Create dataset from spans** | Build a dataset from trace spans |
| | **Append spans to dataset** | Add spans from traces to an existing dataset |
| **Iterate on prompts** | **Create or edit prompt** | Modify the prompt in the playground |
| | **Optimize or align prompt** | Improve prompt based on eval performance or goals |
| | **Load / save prompts** | Load from hub, save to hub, save new versions |
| **Assess quality** | **Build eval** | Write a custom eval for your application |
| | **Choose evals** | Select or attach evals; list evals attached to datasets |
| **Run & compare** | **Run experiment** | Execute an experiment; get results and compare runs |
| | **Get experiments** | Fetch experiment data for analysis and comparison |
### Example prompts
* "Attach the regression-test dataset and use only the question and answer columns"
* "Create a synthetic dataset with 100 examples for regression testing"
* "Create a synthetic dataset with 50 examples that test edge cases"
* "Optimize this prompt to reduce hallucinations"
* "Save this prompt to the hub as a new version"
* "Build an eval to check if the response is helpful"
* "Which experiment has the best hallucination score?"
* "Summarize the eval results across these experiments"
* "Run an experiment and compare the results to the last run"
* "Analyze and compare the outputs of my last two experiments"
* "What's the difference in token usage between experiment A and B?"
Setting up an eval means picking a template, mapping variables, and wiring a data source before you've even started thinking about what good looks like. Alyx understands your project or experiment and builds the eval for you, and even maps variables in a way that makes sense for your data.
### Where to find Alyx
In the eval building, and from the trace slideover or trace page when you want to turn a bad output into a guardrail without starting from a blank template.
### Skills
| Theme | Skill | Description |
| ------------------------- | ------------------------------- | ----------------------------------------------------------- |
| **Target the right data** | **List datasets / experiments** | List datasets and projects for task targeting |
| | **Dataset preview** | Inspect structure and sample rows to choose columns |
| **Build eval** | **Build eval** | Write an eval based on your goals and data structure |
| | **Create / update eval form** | Configure eval parameters and columns |
| **Use eval** | **Configure dataset task** | Set dataset, filters, and sampling for a dataset-based task |
| | **Configure project task** | Set project, filters, and sampling for a trace-based task |
| | **Choose evals** | Select or attach evals |
| | **Propose task name** | Suggest a descriptive name for the task |
### Example prompts
* "Build an eval that checks if the response answers the question"
* "Change this task to use the customer-support dataset"
* "Point this task at the production traces project"
* "What columns should I use for input and output?"
* "Update the sampling rate to 10%"
## Use your own integrations with Alyx
You can connect your own LLM to Alyx. Supported today: **OpenAI** models, **Anthropic** via AWS Bedrock and Vertex AI. More are coming soon. Add and manage integrations in **Settings**, then **Account Settings**, then **Integrations**. For setup details see [AI provider integrations](/ax/security-and-settings/integrations-playground/overview).
## Data privacy
Alyx is built on Arize-hosted models (Azure OpenAI and Claude) for their security and compliance features, keeping your data protected and away from third-party providers.
**Data processing** - Azure OpenAI and Anthropic act as data processors for prompts and outputs sent to and generated by Alyx, depending on which Arize-hosted model is used. The models are stateless, meaning no prompts or outputs are stored.
**No data sharing or model improvement** - Your inputs and outputs are not used to improve OpenAI models, any Microsoft or third-party products, Azure OpenAI models, or Anthropic models.
**Microsoft and Azure OpenAI** - The Azure OpenAI Service is fully controlled by Microsoft and hosted in Microsoft's Azure environment. It does not interact with any other OpenAI-operated services such as ChatGPT or the OpenAI API.
**Anthropic and Claude** - When Alyx uses Claude, Anthropic processes requests on Anthropic infrastructure under Anthropic's security and data commitments for enterprise API use, separate from unrelated consumer products.
**Security and compliance** - Azure OpenAI and Anthropic help Arize meet industry-standard security and compliance measures throughout the process.
For more detail see [Azure OpenAI Service data, privacy, and security](https://learn.microsoft.com/en-us/legal/cognitive-services/openai/data-privacy?tabs=azure-portal) in Microsoft's documentation, [Anthropic's trust and privacy documentation](https://trust.anthropic.com/), or contact [support@arize.com](mailto:support@arize.com).
### Third-party integrations
Alyx includes a support skill that answers user questions. When you ask a support-related question, that question is sent to RunLLM for processing.
Only the specific question you ask is shared with RunLLM. No additional model information or user data is included. You retain control over your interaction with this skill and can revoke consent at any time by contacting [support@arize.com](mailto:support@arize.com). Before using the support skill for the first time you will be asked to acknowledge a one-time disclaimer outlining RunLLM's involvement.