search_deep

A thorough web search for a hard question: up to ten pages, each with the passages that answer it. ₹0.57 per call.

What it does

Send two or three short search_queries and the question itself as objective. data[] holds up to ten pages, each with url, title, publish_date and excerpts[]: passages of the page in Markdown, chosen to answer the objective.

Use it when you want the passages rather than a list of links. For Google's order in a market use search_web, to describe a kind of page use search_semantic, for a written answer use answer_question, and when you already know the page use read_page.

India: there is no country setting. Name the place in the query or the objective.

Inputs

Generated from the contract captured from staging on 2026-11-05.

FieldTypeRequiredDescription
modebasic or advancedno"advanced" (the default) takes about 3 seconds and looks harder; "basic" about 1 second.
objectivestringnoThe question behind the search, in plain words. It focuses the excerpts, so send it.
search_queriesarray of stringyesOne to five short keyword queries, three to six words each. Two or three give the best results.

What drives the cost

Nothing you can turn: flat per call whatever the number of queries (up to five). It always returns pages, even for a nonsense query, so a call is charged unless the source fails. Judge the excerpts against your objective.

Price today

Price today₹0.57 per call (cost_paise: 57) at ₹93.93 per US dollar on 17 November 2026
Hold before the call₹0.86 (hold_paise: 86); when the call settles the real price is charged and the unused part is released
₹0 whenit fails, or the provider answers but there is nothing in it (an empty list, no email, no page): class: "ok_free", ₹0
Timeout30 s, then upstream_timeout (₹0, retryable)
Over MCPrun_tool with tool: "search_deep"
Over RESTPOST https://api.parel.si/v1/run with "tool": "search_deep"
Confirmationnot needed (the estimate is under the ₹200 threshold)

Request

curl https://api.parel.si/v1/run \
  -H "Authorization: Bearer $PAREL_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tool":"search_deep","input":{"objective":"Which Indian D2C skincare brands raised funding in 2026?","search_queries":["Indian D2C skincare funding 2026","skincare brand raises India"]}}'

What comes back

data[], a list of pages. A real answer recorded during the live check of 2026-11-05, trimmed to two pages:

[
  {
    "url": "https://www.bwdisrupt.com/article/asaya-raises-rs-88-cr-at-rs-400-cr-valuation-to-expand-science-backed-skincare-portfolio-620473",
    "title": "Asaya Raises Rs 88 Cr At Rs 400 Cr Valuation To Expand Science-Backed Skincare Portfolio - BW Disrupt",
    "publish_date": "2026-09-17",
    "excerpts": [
      "# Asaya Raises Rs 88 Cr At Rs 400 Cr Valuation To Expand Science-Backed Skincare Portfolio\n\nBW Online Bureau Aug 24, 2026\n\n* [\\# Asaya](https://www.bwdisrupt.com/tag/Asaya)\n\n...\n\n* [\\# RPSG Capital](https://www.bwdisrupt.com/tag/%20RPSG%20Capital)\n\nBengaluru-based Asaya secures USD 9.2 million from …"
    ]
  },
  {
    "url": "https://ascendants.in/funding-feed/asaya-raises-rs-88-crore-d2c-skincare/",
    "title": "Asaya Raises Rs 88 Crore To Expand Its D2C Skincare Business",
    "publish_date": "2026-08-24",
    "excerpts": [
      "D2C skincare brand **[Asaya](https://www.linkedin.com/company/worldofasaya)** has raised Rs 88 crore in a new funding round led by RPSG Capital, with participation from OTP Ventures, Huddle Ventures, Hyperscale Ventures and 72 Ventures.\n\n...\n\nThe brand has also developed products intended to address…"
    ]
  }
]

Notes

  • Flat price per call; up to ten pages and five queries.
  • Always answers with pages, even for a nonsense query, so check that the excerpts answer the objective.
  • No country setting: name the place in the query.

Siblings

This page as Markdown: /docs/tools/search_deep.md. Something wrong or missing: hello@parel.si.