# Web Search Tool

The Web Search Tool allows DinoAI to search the internet and extract content from specific URLs, bringing external documentation, package references, and web resources directly into your development workflow without leaving the Code IDE.

This tool gives DinoAI access to up-to-date information beyond its training data — from dbt package docs to Paradime release notes — and lets it read any URL you point it to for deeper context.

### Capabilities

The Web Search Tool exposes two underlying operations:

**`execute_web_search`** performs a web search using a concise query and returns structured results including titles, URLs, and summaries. You can optionally restrict results to specific domains — such as `docs.getdbt.com` or `docs.paradime.io/app-help` — to keep answers focused and relevant.

**`extract_url_content`** fetches and extracts the full text content of a specific URL. Use this when you have a particular page in mind and want DinoAI to read, summarize, or reason over its contents directly.

### Using the Web Search Tool

1. Open DinoAI in the right panel of the Code IDE
2. Ask a question or provide a URL you want DinoAI to look up
3. DinoAI will search the web or extract the page content automatically — no permission prompt is required
4. Review the information DinoAI surfaces and incorporate it into your implementation

{% hint style="info" %}
Web Search is enabled by default and does not require a separate integration setup.
{% endhint %}

### Example Use Cases

#### Looking Up dbt Documentation

**Prompt**

```
What arguments does the dbt source freshness command accept?
```

**Result:** DinoAI searches `docs.getdbt.com`, retrieves the relevant page, and summarizes the available CLI arguments — so you can run the right command without switching tabs.

***

#### Finding a dbt Package

**Prompt**

```
Is there a dbt package for asserting column uniqueness across tables?
```

**Result:** DinoAI searches `hub.getdbt.com` and surfaces matching packages with descriptions and links, helping you evaluate options before adding a dependency to your project.

***

#### Reading a Specific URL

**Prompt**

```
Read https://docs.paradime.io/app-help/documentation/dino-ai/tools-and-features/notion-tool and summarize how the Notion Tool works.
```

**Result:** DinoAI fetches the page, extracts its content, and returns a structured summary — useful for quickly onboarding to a new tool or verifying integration setup steps.

***

#### Checking Paradime Release Notes

**Prompt**

```
Search the Paradime docs for what changed in the latest DinoAI release.
```

**Result:** DinoAI queries `docs.paradime.io/app-help` and surfaces the most relevant changelog or release note entries, keeping you current without manual browsing.

### Working with Other Tools

The Web Search Tool works well alongside DinoAI's other capabilities:

* Combine with the **Terminal Tool** to look up a command's syntax and run it immediately in your project environment
* Use alongside **Notion** or **Confluence** tools to cross-reference internal specs against external documentation
* Pair with **Linear** or **Jira** tools to pull in ticket context while researching the right implementation approach online

### Best Practices

**Be specific in your query** — Shorter, precise search queries (under 300 characters) return more relevant results than broad, open-ended questions.

**Use domain restrictions for focused searches** — When you know the answer lives in a specific site (e.g. `docs.getdbt.com`), tell DinoAI to restrict the search there. This reduces noise and speeds up the response.

Useful domains to reference:

* `docs.paradime.io/app-help` — Paradime product documentation
* `docs.getdbt.com` — dbt Core documentation
* `hub.getdbt.com` — dbt package registry
* `elementary-data.com/dbt-test-hub` — Elementary test catalogue

**Provide URLs directly when you have them** — If you already know the page you want DinoAI to read, use `extract_url_content` by sharing the URL in your prompt. This is faster and more precise than a keyword search.

**Verify before implementing** — Web search results reflect the state of the web at the time of the query and may include outdated or irrelevant content. Always review DinoAI's interpretation before applying it to your project.
