POST
/
desearch
/
ai
/
search
curl --request POST \
  --url https://apis.datura.ai/desearch/ai/search \
  --header 'Authorization: <api-key>' \
  --header 'Content-Type: application/json' \
  --data '{
  "prompt": "Bittensor",
  "tools": [
    "web",
    "hackernews",
    "reddit",
    "wikipedia",
    "youtube",
    "twitter",
    "arxiv"
  ],
  "model": "NOVA",
  "date_filter": "PAST_24_HOURS",
  "streaming": true,
  "result_type": "LINKS_WITH_FINAL_SUMMARY",
  "system_message": "Summarize the content by categorizing key points into '\''Pros'\'' and '\''Cons'\'' sections."
}'
{
  "wikipedia_search_results": [
    {
      "organic_results": [
        {
          "title": "Example Title",
          "url": "https://example.com",
          "link": "https://example.com/link",
          "snippet": "This is an example snippet from the search result.",
          "summary_description": "This is a summary description of the search result."
        }
      ]
    }
  ],
  "youtube_search_results": [
    {
      "organic_results": [
        {
          "title": "Example Title",
          "url": "https://example.com",
          "link": "https://example.com/link",
          "snippet": "This is an example snippet from the search result.",
          "summary_description": "This is a summary description of the search result."
        }
      ]
    }
  ],
  "arxiv_search_results": [
    {
      "organic_results": [
        {
          "title": "Example Title",
          "url": "https://example.com",
          "link": "https://example.com/link",
          "snippet": "This is an example snippet from the search result.",
          "summary_description": "This is a summary description of the search result."
        }
      ]
    }
  ],
  "reddit_search_results": [
    {
      "organic_results": [
        {
          "title": "Example Title",
          "url": "https://example.com",
          "link": "https://example.com/link",
          "snippet": "This is an example snippet from the search result.",
          "summary_description": "This is a summary description of the search result."
        }
      ]
    }
  ],
  "hacker_news_search_results": [
    {
      "organic_results": [
        {
          "title": "Example Title",
          "url": "https://example.com",
          "link": "https://example.com/link",
          "snippet": "This is an example snippet from the search result.",
          "summary_description": "This is a summary description of the search result."
        }
      ]
    }
  ],
  "text_chunks": [
    {
      "twitter_summary": [
        "<string>"
      ]
    }
  ],
  "completion_links": [
    "<string>"
  ],
  "search_completion_links": [
    "<string>"
  ],
  "completion": {
    "key_tweets": [
      {
        "text": "This is an example tweet text.",
        "url": "https://twitter.com/example_tweet"
      }
    ],
    "twitter_summary": "This is an example Twitter summary.",
    "summary": "This is an example summary."
  },
  "miner_link_scores": {}
}

Authorizations

​
Authorization
string
header
required

Body

application/json
​
prompt
string
required

Search query prompt

Example:

"Bittensor"

​
tools
(enum<string> | string)[]
required

A list of tools to be used for the search

Available options:
web,
hackernews,
reddit,
wikipedia,
youtube,
twitter,
arxiv
Example:
[
  "web",
  "hackernews",
  "reddit",
  "wikipedia",
  "youtube",
  "twitter",
  "arxiv"
]
​
model
enum<string>
required

The model to be used for the search

Available options:
NOVA,
ORBIT,
HORIZON
​
date_filter
enum<string> | null
required

The date filter to be used for the search

Available options:
PAST_24_HOURS,
PAST_2_DAYS,
PAST_WEEK,
PAST_2_WEEKS,
PAST_MONTH,
PAST_2_MONTHS,
PAST_YEAR,
PAST_2_YEARS
Example:

"PAST_24_HOURS"

​
streaming
boolean
required

Whether to stream results

Example:

true

​
result_type
enum<string> | null

The result type to be used for the search

Available options:
ONLY_LINKS,
LINKS_WITH_SUMMARIES,
LINKS_WITH_FINAL_SUMMARY
Example:

"LINKS_WITH_FINAL_SUMMARY"

​
system_message
string | null

The system message to be used for the search

Example:

"Summarize the content by categorizing key points into 'Pros' and 'Cons' sections."

Response

200
application/json
A JSON object mapping tool names to their search results.
​
wikipedia_search_results

Search results retrieved from Wikipedia, containing relevant articles based on the query.

​
youtube_search_results

Search results from YouTube, including relevant videos matching the search criteria.

​
arxiv_search_results

Search results from the Arxiv repository, returning academic papers related to the query.

​
reddit_search_results

Search results from Reddit, providing discussions and posts related to the search query.

​
hacker_news_search_results

Search results from Hacker News, listing articles and discussions relevant to the query.

​
text_chunks

Extracted or summarized text chunks related to the search query, which may be in list or string format.

​
completion_links
string[] | null

A list of generated links related to the search query for further exploration.

​
search_completion_links
string[] | null

Generated links specifically related to search completion suggestions or follow-up queries.

​
completion
object | null

Generated completion text or response, typically used in AI-generated content results.

​
miner_link_scores
object | null

A list of miner link scores, typically used in AI-generated content results.