exa-haystack

Haystack integration for Exa - the search engine designed for AI.

Installation

pip install exa-haystack

Usage

ExaWebSearch

Search the web using Exa's neural search:

from haystack_integrations.components.websearch.exa import ExaWebSearch

search = ExaWebSearch(num_results=5)
results = search.run(query="latest developments in quantum computing")

for doc in results["documents"]:
    print(f"Title: {doc.meta['title']}")
    print(f"URL: {doc.meta['url']}")
    print(f"Content: {doc.content[:200]}...")
    print()

ExaFindSimilar

Find pages similar to a given URL:

from haystack_integrations.components.websearch.exa import ExaFindSimilar

find_similar = ExaFindSimilar(num_results=5)
results = find_similar.run(url="https://example.com/article")

for doc in results["documents"]:
    print(f"Similar: {doc.meta['title']} - {doc.meta['url']}")

ExaContents

Fetch full content for URLs:

from haystack_integrations.components.websearch.exa import ExaContents

contents = ExaContents(text=True, highlights=True, summary=True)
results = contents.run(urls=["https://example.com/article1", "https://example.com/article2"])

for doc in results["documents"]:
    print(f"Title: {doc.meta['title']}")
    print(f"Summary: {doc.meta.get('summary', 'N/A')}")
    print(f"Content: {doc.content[:500]}...")

Using in a Pipeline

from haystack import Pipeline
from haystack.components.builders import PromptBuilder
from haystack.components.generators import OpenAIGenerator
from haystack_integrations.components.websearch.exa import ExaWebSearch

template = """
Based on the following search results, answer the question.

Search Results:
{% for doc in documents %}
Title: {{ doc.meta.title }}
Content: {{ doc.content }}
---
{% endfor %}

Question: {{ query }}
Answer:
"""

pipe = Pipeline()
pipe.add_component("search", ExaWebSearch(num_results=3))
pipe.add_component("prompt", PromptBuilder(template=template))
pipe.add_component("llm", OpenAIGenerator())

pipe.connect("search.documents", "prompt.documents")
pipe.connect("prompt", "llm")

result = pipe.run({
    "search": {"query": "What is retrieval augmented generation?"},
    "prompt": {"query": "What is retrieval augmented generation?"}
})

print(result["llm"]["replies"][0])

Configuration

Set your Exa API key as an environment variable:

export EXA_API_KEY="your-api-key"

Or pass it directly:

from haystack.utils import Secret
from haystack_integrations.components.websearch.exa import ExaWebSearch

search = ExaWebSearch(api_key=Secret.from_token("your-api-key"))

Components

ExaWebSearch

Main search component using Exa's /search endpoint.

Parameters:

• api_key: Exa API key (default: from EXA_API_KEY env var)
• num_results: Number of results to return (default: 10)
• type: Search type — "auto", "neural", "fast", "deep", "deep-reasoning", "deep-max", "instant" (default: "auto")
• include_domains / exclude_domains: Domain filters
• start_published_date / end_published_date: Date range filters (ISO format)
• category: Search category filter ("company", "research paper", "news", "pdf", "tweet", "personal site", "financial report", "people")
• additional_queries: Extra queries for deep search
• output_schema: JSON schema for structured deep search output (returned as deep_output component output)
• max_age_hours: Content freshness control (0 = always livecrawl, -1 = cache only, positive = max cache age in hours)
• livecrawl: Legacy freshness option (deprecated, use max_age_hours)
• text / highlights / summary: Content options (bool or dict with sub-options)

ExaFindSimilar

Find similar pages using Exa's /findSimilar endpoint.

Parameters:

• api_key: Exa API key (default: from EXA_API_KEY env var)
• num_results: Number of results to return (default: 10)
• include_domains / exclude_domains: Domain filters
• exclude_source_domain: Exclude the source URL's domain from results
• category: Search category filter
• max_age_hours: Content freshness control
• livecrawl: Legacy freshness option (deprecated, use max_age_hours)
• text / highlights / summary: Content options

ExaContents

Fetch content for URLs using Exa's /contents endpoint.

Parameters:

• api_key: Exa API key (default: from EXA_API_KEY env var)
• text: Include full text content (default: True)
• highlights: Include highlighted snippets
• summary: Include AI-generated summary
• max_age_hours: Content freshness control
• livecrawl: Legacy freshness option (deprecated, use max_age_hours)
• subpages / subpage_target: Crawl linked subpages
• extras / flags: Additional extraction options

Outputs: documents (list of Documents) and statuses (per-URL status/error info)

ExaAnswer

Generate AI-powered answers with citations using Exa's /answer endpoint.

Parameters:

• api_key: Exa API key (default: from EXA_API_KEY env var)
• model: "exa" or "exa-pro" (default: "exa")
• system_prompt: Custom system prompt
• output_schema: JSON schema for structured answer output

ExaStreamAnswer

Streaming variant of ExaAnswer — yields answer chunks via SSE.

ExaResearch

Conduct deep research using Exa's /research/v1 endpoint.

Parameters:

• api_key: Exa API key (default: from EXA_API_KEY env var)
• model: "exa-research-fast", "exa-research", or "exa-research-pro" (default: "exa-research")
• output_schema: JSON schema for structured research output
• poll_interval / max_wait_time: Polling configuration

License

Apache-2.0