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

# Set up

> Choose your category or job, tools, frameworks, requests and models.

For a new project, choose **Agent Picks** during setup and enter your product name and domain. For an existing project, open **Agent Picks** and select **Set up Agent Picks**.

To review or change a saved setup, open the project's **Settings › Agent Picks**. Only members who can manage the project can save changes.

## Covered category

1. Choose **Use a public category** as your starting point.
2. Select the category and your tool. If your tool is absent, choose **Your tool isn't listed**. Results can show what agents pick, but cannot measure your tool until it is recognized in the test.
3. Select up to six competitors.
4. Choose the projects agents start in, then the agents and models.
5. Select **Save setup**.

A category supplies its tools and requests. Requests ask agents to add an integration, recommend a tool, or use a hosted service. Categories without a hosted option use the first two requests.

If your project has no finished test, Agent Picks can show the category's public test. The page labels these results **Public test**. They are not a private test of your project.

## Your own job

Choose **Start from scratch** to define a job instead of using the category unchanged. If your product has no covered category, setup opens the job editor. Starting from a covered category fills in its requests and tools for you to edit.

1. Describe the work in one sentence, such as "Add AI traffic analytics to this app." This becomes the build request.
2. Review the drafted recommendation request and, when relevant, the hosted-service request. Edit or reset them before saving. Use plain requests without links, code or environment variable names.
3. Confirm your tool and the competitors to compare.
4. Review how DevTune recognizes each tool: its packages, environment names, API hosts, code patterns and domains. Use **Draft again** or correct the fields yourself if a draft is missing or wrong.
5. Review **How we decide Built its own**. Leave fields empty when they do not apply. Ask a teammate familiar with your integration to check package names and code patterns if needed.
6. Choose frameworks and models, then save.

If you change the job sentence, redraft the requests or select **Confirm requests** before saving. A saved job compares with its own earlier tests and has no public test.

<Note>
  If the app says runs for a custom job are unavailable, your setup stays saved.
  You can run its starter test from Agent Picks when the app offers it.
</Note>

## Frameworks

Choose environments that match where buyers use your product. Options include an empty project, Next.js, TanStack Start, SvelteKit, Express, FastAPI and Rails. Some environments vary the starting state, such as Next.js with an existing tool.

The starter test uses one plain framework starter, or the empty project. During onboarding, mark the framework to use for the starter test. Paid tests use the saved environment selections.

## Agents and models

Paid tests offer these models:

| Agent | Models |
| - | - |
| Claude Code | Claude Sonnet 5, Claude Opus 5.5, Claude Fable 5.1, Claude Haiku 5.5 |
| Codex | GPT-6.1 Sol, GPT-6 Astra, GPT-6 Luna |

The defaults are Claude Sonnet 5 and GPT-6.1 Sol. Selecting two models for one agent tests both separately and increases the run count. These choices do not change the starter test's fixed models.

Save changes before [requesting a paid test](/agent-picks/paid-tests-and-credits). If the app warns that saved models are unavailable, review the current choices and save again.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.