Parallel

POST /search/parallel, forwarded to POST https://api.parallel.ai/v1/search.

Request

curl -X POST "https://api.withgrowl.com/v1/search/parallel" \
  -H "Authorization: Bearer elo_pub_..." \
  -H "Content-Type: application/json" \
  -d '{
    "objective": "travel to vietnam",
    "search_queries": ["travel to vietnam"],
    "advanced_settings": {
      "max_results": 5,
      "source_policy": {
        "include_domains": ["vietnam.travel"],
        "after_date": "2026-02-01"
      }
    }
  }'

Common fields are below. The body goes to Parallel as sent, so everything in Parallel’s reference works; that reference is the complete, current schema.

Field	Description
objective	What you are trying to find, in plain language.
search_queries	One or more queries to run for it.
advanced_settings.max_results	Number of results.
advanced_settings.location	Country code, lower case.
advanced_settings.source_policy.include_domains	Only these hostnames.
advanced_settings.source_policy.exclude_domains	Never these hostnames.
advanced_settings.source_policy.after_date	YYYY-MM-DD. Only pages published after it.

Response

Results are under results. Each has title, url, excerpts (a list of passages) and publish_date.

{
  "results": [
    {
      "title": "Visit Vietnam: The Official Tourism Website of Vietnam",
      "url": "https://vietnam.travel/",
      "excerpts": [
        "Discover the highlights of Vietnam through the country's official tourism website."
      ],
      "publish_date": "2026-02-18"
    }
  ]
}

Through Elo

• Parallel takes its key in x-api-key. Keep that header and put your Elo key in it, or use Authorization: Bearer; Elo sets Parallel’s key itself.
• A search with several search_queries is one call to Elo and one call to Parallel; Parallel prices it by its own rules.