1  ---
  2  title: "SerperDevWebSearch"
  3  id: serperdevwebsearch
  4  slug: "/serperdevwebsearch"
  5  description: "Search engine using SerperDev API."
  6  ---
  7  
  8  # SerperDevWebSearch
  9  
 10  Search engine using SerperDev API.
 11  
 12  <div className="key-value-table">
 13  
 14  |  |  |
 15  | --- | --- |
 16  | **Most common position in a pipeline** | Before [`LinkContentFetcher`](../fetchers/linkcontentfetcher.mdx)  or [Converters](../converters.mdx) |
 17  | **Mandatory init variables** | `api_key`: The SearchAPI API key. Can be set with `SERPERDEV_API_KEY` env var. |
 18  | **Mandatory run variables** | `query`: A string with your query |
 19  | **Output variables** | `documents`: A list of documents  <br /> <br />`links`: A list of strings of resulting links |
 20  | **API reference** | [Websearch](/reference/websearch-api) |
 21  | **GitHub link** | https://github.com/deepset-ai/haystack/blob/main/haystack/components/websearch/serper_dev.py |
 22  
 23  </div>
 24  
 25  ## Overview
 26  
 27  When you give `SerperDevWebSearch` a query, it returns a list of the URLs most relevant to your search. It uses page snippets (pieces of text displayed under the page title in search results) to find the answers, not the whole pages.
 28  
 29  To search the content of the web pages, use the [`LinkContentFetcher`](../fetchers/linkcontentfetcher.mdx) component.
 30  
 31  `SerperDevWebSearch` requires a [SerperDev](https://serper.dev/) key to work. It uses a `SERPERDEV_API_KEY` environment variable by default. Otherwise, you can pass an `api_key` at initialization – see code examples below.
 32  
 33  :::info[Alternative search]
 34  
 35  To use [Search API](https://www.searchapi.io/) as an alternative, see its respective [documentation page](searchapiwebsearch.mdx).
 36  :::
 37  
 38  ## Usage
 39  
 40  ### On its own
 41  
 42  This is an example of how `SerperDevWebSearch` looks up answers to our query on the web and converts the results into a list of documents with content snippets of the results, as well as URLs as strings.
 43  
 44  ```python
 45  from haystack.components.websearch import SerperDevWebSearch
 46  from haystack.utils import Secret
 47  
 48  web_search = SerperDevWebSearch(api_key=Secret.from_token("<your-api-key>"))
 49  query = "What is the capital of Germany?"
 50  
 51  response = web_search.run(query)
 52  ```
 53  
 54  ### In a pipeline
 55  
 56  Here’s an example of a RAG pipeline where we use a `SerperDevWebSearch` to look up the answer to the query. The resulting documents are then passed to `LinkContentFetcher` to get the full text from the URLs. Finally, `PromptBuilder` and `OpenAIGenerator` work together to form the final answer.
 57  
 58  ```python
 59  from haystack import Pipeline
 60  from haystack.utils import Secret
 61  from haystack.components.builders.chat_prompt_builder import ChatPromptBuilder
 62  from haystack.components.fetchers import LinkContentFetcher
 63  from haystack.components.converters import HTMLToDocument
 64  from haystack.components.generators.chat import OpenAIChatGenerator
 65  from haystack.components.websearch import SerperDevWebSearch
 66  from haystack.dataclasses import ChatMessage
 67  from haystack.utils import Secret
 68  
 69  web_search = SerperDevWebSearch(api_key=Secret.from_token("<your-api-key>"), top_k=2)
 70  link_content = LinkContentFetcher()
 71  html_converter = HTMLToDocument()
 72  
 73  prompt_template = [
 74      ChatMessage.from_system("You are a helpful assistant."),
 75      ChatMessage.from_user(
 76          "Given the information below:\n"
 77          "{% for document in documents %}{{ document.content }}{% endfor %}\n"
 78          "Answer question: {{ query }}.\nAnswer:",
 79      ),
 80  ]
 81  
 82  prompt_builder = ChatPromptBuilder(
 83      template=prompt_template,
 84      required_variables={"query", "documents"},
 85  )
 86  llm = OpenAIChatGenerator(
 87      api_key=Secret.from_token("<your-api-key>"),
 88      model="gpt-3.5-turbo",
 89  )
 90  
 91  pipe = Pipeline()
 92  pipe.add_component("search", web_search)
 93  pipe.add_component("fetcher", link_content)
 94  pipe.add_component("converter", html_converter)
 95  pipe.add_component("prompt_builder", prompt_builder)
 96  pipe.add_component("llm", llm)
 97  
 98  pipe.connect("search.links", "fetcher.urls")
 99  pipe.connect("fetcher.streams", "converter.sources")
100  pipe.connect("converter.documents", "prompt_builder.documents")
101  pipe.connect("prompt_builder.messages", "llm.messages")
102  
103  query = "What is the most famous landmark in Berlin?"
104  
105  pipe.run(data={"search": {"query": query}, "prompt_builder": {"query": query}})
106  ```
107  
108  ## Additional References
109  
110  :notebook: Tutorial: [Building Fallbacks to Websearch with Conditional Routing](https://haystack.deepset.ai/tutorials/36_building_fallbacks_with_conditional_routing)