Octen Broad Search for Research and Comparisons

Give one research question. Octen develops related queries from several angles and searches them concurrently, returning sources grouped by query. Your Agent can compare the material, remove duplicate references and check sources before drawing conclusions.

Your AI

Use with your AI

For tasks with several steps, continue in the AI you already use.

CodexClaude CodeCursorWindsurfCline

View instructions

Connect to Toolmu using https://toolmu.com/SKILL.md if needed. Find octen-broad-search-octen and check its current availability, inputs and price against my authorized budget before running the task. Ask for my research question if it is missing, and put it in body.query. Use body.max_queries and body.search_options only to match the scope I request; distinguish the number of generated searches from the requested results per search. Each executed sub-query is a billable search unit, so choose a useful scope within my spending authority. Apply any requested topic, domain, text, country or time preferences consistently across the sub-queries. Use either a relative time range or explicit time boundaries, and keep publication dates distinct from crawl dates. Review the generated queries and the actual references in each group. Preserve available titles, links, source wording and dates, and make missing or truncated material explicit. Compare the groups, identify repeated references and keep citations attached to the source claims. A highlight or bounded page text is not proof of a complete article. Formatting, source opening, fact checking and a final comparison are my Agent tasks after retrieval; do not claim the search itself has verified every statement. Report the useful material, unresolved gaps and actual execution charge when relevant.

Task:
Search for [cloud GPU services to compare prices, available hardware, and service locations]. Use up to [4 related questions] and show up to [3 results for each question]. Group the links by question, remove duplicates, and point out important claims to verify.

Paste into your connected AI's conversation.

Connection guide

Up to 360.036 credits / call

The final charge is based on confirmed usage. Unused reserved credits return to your available balance.

Octen

Octen Broad Search for Research and Comparisons

You receive

Receive generated search queries and the references grouped under each. Available results can include page titles, links and relevant excerpts, with optional page text or images when requested and available. Your Agent can arrange the groups into a source comparison, keeping each reference attached to the angle it addresses.

Start with your task

Search for [cloud GPU services to compare prices, available hardware, and service locations]. Use up to [4 related questions] and show up to [3 results for each question]. Group the links by question, remove duplicates, and point out important claims to verify.

What to know

Search groups are research material, not a finished or verified report. More queries do not prove exhaustive coverage or unique sources. Requested result counts and text limits do not guarantee enough accessible content, and recent search or crawl activity does not establish that a claim is current.

Inputs & outputs

Inspect the current inputs, availability, and maximum charge before starting.

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Research request`body`	`object`	Required	Required object containing the original question and optional sub-query and search settings. Required fields `query`
Original question`body.query`	`string`	Required	Required search question, 1–500 characters. Octen uses it to generate related sub-queries. Minimum length `1` Maximum length `500`
Sub-query cap`body.max_queries`	`integer`	Optional	Upper bound on generated sub-queries: integer 1–30, default 5. Each executed sub-query is a billable search unit; this is not a guaranteed number of angles. Default `5` Minimum `1` Maximum `30`
Shared search settings`body.search_options`	`object`	Optional	Optional search settings applied to every generated sub-query, including result count, content and source preferences.
Results per sub-query`body.search_options.count`	`integer`	Optional	Requested results for each search: integer 1–100, default 5. This is separate from the sub-query cap and is not a guarantee of available results. Default `5` Minimum `1` Maximum `100`
Search topic`body.search_options.topic`	`string`	Optional	general for general web search or news for news-focused results; default general. Default `general` Options `general, news`
Highlight format`body.search_options.format`	`string`	Optional	Format of highlights: text or markdown; default text. This does not request a finished Markdown report. Default `text` Options `markdown, text`
Country preference`body.search_options.country`	`string`	Optional	ISO 3166 country preference for prioritizing results. When omitted, Octen infers a preference from the question.
End time`body.search_options.end_time`	`string`	Optional	End boundary for result time filtering, expressed as an ISO 8601 string. Use with the selected time basis rather than a relative range.
Relevant excerpt settings`body.search_options.highlight`	`object`	Optional	Optional settings for requesting query-relevant excerpts and choosing their token limit.
Request highlights`body.search_options.highlight.enable`	`boolean`	Optional	Boolean, default true. Request a relevant highlight with each result when available. Default `true`
Highlight token cap`body.search_options.highlight.max_tokens`	`integer`	Optional	Maximum tokens per highlight: integer 100–20,000, default 512. Tokens are not a character or word count. Default `512` Minimum `100` Maximum `20000`
Explicit-content filter`body.search_options.safesearch`	`string`	Optional	strict or off for explicit/adult-content filtering; default strict. Default `strict` Options `off, strict`
Start time`body.search_options.start_time`	`string`	Optional	Start boundary for result time filtering, expressed as an ISO 8601 string. Use with the selected time basis rather than a relative range.
Time-filter basis`body.search_options.time_basis`	`string`	Optional	auto, published or crawled; default auto. Select whether filtering uses publication or crawl time; these dates have different meanings. Default `auto` Options `auto, published, crawled`
Relative time window`body.search_options.time_range`	`string`	Optional	day/week/month/year or d/w/m/y, counting backward from now. Use this or start_time/end_time; explicit boundaries take precedence. Options `day, week, month, year, d, w, m, y`
Excluded page text`body.search_options.exclude_text`	`array`	Optional	Up to 5 text strings that must not appear in result-page text; each string is at most 30 characters. Maximum items `5`
Excluded text item`body.search_options.exclude_text[]`	`string`	—	One text string to exclude from result-page text, at most 30 characters. Maximum length `30`
Longer page-text settings`body.search_options.full_content`	`object`	Optional	Optional settings for requesting page text and limiting tokens per result. A text budget is not proof of an entire page.
Request page text`body.search_options.full_content.enable`	`boolean`	Optional	Boolean, default false. Request full_content when available, subject to its token cap. Default `false`
Page-text token cap`body.search_options.full_content.max_tokens`	`integer`	Optional	Maximum tokens of page text per result: integer 100–100,000, default 2,048. Returned text can be truncated. Default `2048` Minimum `100` Maximum `100000`
Required page text`body.search_options.include_text`	`array`	Optional	Up to 5 strings that must appear in result-page text; each string is at most 30 characters. Maximum items `5`
Required text item`body.search_options.include_text[]`	`string`	—	One required string in result-page text, at most 30 characters. Maximum length `30`
Request images`body.search_options.include_images`	`boolean`	Optional	Boolean, default false. Request available images with the search results; this does not save media files. Default `false`
Excluded domains`body.search_options.exclude_domains`	`array`	Optional	Domains excluded from results: at most 1,200 strings, each at most 60 characters. Maximum items `1200`
Excluded domain item`body.search_options.exclude_domains[]`	`string`	—	One domain string to exclude, at most 60 characters. Maximum length `60`
Included domains`body.search_options.include_domains`	`array`	Optional	Domains to include in the search: at most 1,200 strings, each at most 60 characters. Maximum items `1200`
Included domain item`body.search_options.include_domains[]`	`string`	—	One domain string to include, at most 60 characters. Maximum length `60`

View raw JSON

{
  "type": "object",
  "required": [
    "body"
  ],
  "properties": {
    "body": {
      "type": "object",
      "$schema": "https://json-schema.org/draft/2020-12/schema",
      "required": [
        "query"
      ],
      "~standard": {
        "vendor": "zod",
        "version": 1,
        "jsonSchema": {}
      },
      "properties": {
        "query": {
          "type": "string",
          "maxLength": 500,
          "minLength": 1,
          "description": "The original search query."
        },
        "max_queries": {
          "type": "integer",
          "default": 5,
          "maximum": 30,
          "minimum": 1,
          "description": "Upper bound on the number of sub-queries generated (each executed sub-query is a billed search unit)."
        },
        "search_options": {
          "type": "object",
          "properties": {
            "count": {
              "type": "integer",
              "default": 5,
              "maximum": 100,
              "minimum": 1,
              "description": "Number of results to return (1-100)."
            },
            "topic": {
              "enum": [
                "general",
                "news"
              ],
              "type": "string",
              "default": "general",
              "description": "Use 'general' for general web search, or 'news' for news-focused results."
            },
            "format": {
              "enum": [
                "markdown",
                "text"
              ],
              "type": "string",
              "default": "text",
              "description": "Formatting of highlight outputs."
            },
            "country": {
              "type": "string",
              "description": "ISO 3166 country to prioritize results from; default 'auto' infers it from the query."
            },
            "end_time": {
              "type": "string",
              "description": "End time for filtering results. ISO 8601 format."
            },
            "highlight": {
              "type": "object",
              "properties": {
                "enable": {
                  "type": "boolean",
                  "default": true,
                  "description": "If true, returns query-relevant highlight in each result."
                },
                "max_tokens": {
                  "type": "integer",
                  "default": 512,
                  "maximum": 20000,
                  "minimum": 100,
                  "description": "Max tokens returned per highlight."
                }
              },
              "additionalProperties": false
            },
            "safesearch": {
              "enum": [
                "off",
                "strict"
              ],
              "type": "string",
              "default": "strict",
              "description": "Explicit/adult content filtering."
            },
            "start_time": {
              "type": "string",
              "description": "Start time for filtering results. ISO 8601 format."
            },
            "time_basis": {
              "enum": [
                "auto",
                "published",
                "crawled"
              ],
              "type": "string",
              "default": "auto",
              "description": "Which time field is used for time filtering."
            },
            "time_range": {
              "enum": [
                "day",
                "week",
                "month",
                "year",
                "d",
                "w",
                "m",
                "y"
              ],
              "type": "string",
              "description": "Relative time window counting back from now. Mutually exclusive with start_time/end_time (which take precedence)."
            },
            "exclude_text": {
              "type": "array",
              "items": {
                "type": "string",
                "maxLength": 30
              },
              "maxItems": 5,
              "description": "Strings that must not appear in the result page text."
            },
            "full_content": {
              "type": "object",
              "properties": {
                "enable": {
                  "type": "boolean",
                  "default": false,
                  "description": "If true, returns full_content for each result."
                },
                "max_tokens": {
                  "type": "integer",
                  "default": 2048,
                  "maximum": 100000,
                  "minimum": 100,
                  "description": "Maximum tokens of full content included per result."
                }
              },
              "additionalProperties": false
            },
            "include_text": {
              "type": "array",
              "items": {
                "type": "string",
                "maxLength": 30
              },
              "maxItems": 5,
              "description": "Strings that must appear in the result page text."
            },
            "include_images": {
              "type": "boolean",
              "default": false,
              "description": "Whether to include images in each result."
            },
            "exclude_domains": {
              "type": "array",
              "items": {
                "type": "string",
                "maxLength": 60
              },
              "maxItems": 1200,
              "description": "Domains to specifically exclude from the search results."
            },
            "include_domains": {
              "type": "array",
              "items": {
                "type": "string",
                "maxLength": 60
              },
              "maxItems": 1200,
              "description": "Domains to specifically include in the search results."
            }
          },
          "description": "Search options applied to each sub-query (same parameters and defaults as Web Search).",
          "additionalProperties": false
        }
      },
      "additionalProperties": false
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

Octen expands the original question into related queries and searches those angles concurrently. The response groups references by generated query, so you can inspect which parts of the question were explored. The sub-query cap limits the requested search scope; it does not promise coverage of every angle.

Read available titles, destination links and query-relevant excerpts within their groups. When author or publication and crawl dates are supplied, retain that context. A reference can appear under more than one query. Your Agent can deduplicate references while keeping the connections between sources and research angles.

Search options apply to each generated query. Topic, domain and text preferences narrow the requested search, while a country preference prioritizes geographic context. Use the input reference for their limits. These choices do not guarantee a result count or establish that an omitted source does not exist.

Highlights provide relevant excerpts. Optional full content requests longer page text, still subject to the chosen token cap and available material. Text or Markdown formatting applies to highlights; it does not turn the response into a completed report. Requested images are additional search material, rather than a promise of saved media files.

For time filtering, choose a relative range or explicit ISO 8601 boundaries and select the publication or crawl basis that fits the question. A newly crawled page can contain older information. Your Agent should preserve source dates, check important claims against the linked material and distinguish its comparison or synthesis from the returned references.

Octen: Broad Search and grouped research angles

Octen: Broad Search request reference

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

The sub-query cap bounds the generated searches; count requests results within each search. Their defaults are both 5, but neither establishes a unique-source count or complete coverage. Review the actual groups and references.

Octen Broad Search for Research and Comparisons

Octen Broad Search for Research and Comparisons

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Describe what you want to learn

Choose how broad and detailed to search

Review the links and check important claims

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Describe what you want to learn

Choose how broad and detailed to search

Review the links and check important claims

Questions about this tool.

How are sub-queries different from results per query?

Does requesting full content give me complete articles?

Does a recent crawl date mean recently published information?

Will the groups become a finished report with unique citations?