Extract Text from Websites with Octen

Give your Agent 1–20 known page URLs and request their main text as Markdown or plain text. Add a question or keywords to request relevant passages instead. Octen extracts the page material; your Agent can then compare it or prepare source-linked notes.

Your AI

Use with your AI

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

CodexClaude CodeCursorWindsurfCline

View instructions

Use Toolmu's octen-extract-octen, catalog ID 515b466b-0f03-4c5c-ac6a-ccb36846f5ea, to read the page URLs I provide. Reuse the existing Toolmu connection; if setup is needed, follow https://toolmu.com/SKILL.md. Read get_tool for this exact Tool, check current availability and price, and honor my existing authorization and budget. Ask for the URL list if it is missing. Request main content unless I ask for focused highlights; in that case use query for my question. Apply my chosen text format, media references, cache tolerance or per-page time limit when useful. Inspect each URL's actual text and status. Separate readable source content from login screens, errors or missing main content. Preserve source associations and any useful supplied title or dates; keep publication and crawl dates distinct. Prepare the requested notes or comparison as separate Agent work, cite only material you actually received, and identify gaps without inventing explanations. Fetching media files or transcribing them is a separate step.

Task:
Read [page 1 link] and [page 2 link], then compare their [application requirements]. Keep each page’s main text with its source link and tell me which requirements are unclear or missing.

Paste into your connected AI's conversation.

Connection guide

Up to 0.024 credits / call

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

Octen

Extract Text from Websites with Octen

You receive

For pages that supply readable content, the result provides text or relevant highlights associated with each URL, alongside its extraction status. When supplied, titles, page labels, dates and requested media references add context. Your Agent can turn that material into a reading brief or comparison while keeping each point linked to its source.

Start with your task

Read [page 1 link] and [page 2 link], then compare their [application requirements]. Keep each page’s main text with its source link and tell me which requirements are unclear or missing.

What to know

Extraction does not discover a whole website or verify the claims in its text. A fetched login or error page is not a readable article. Source access, available main content and cache settings affect what you can use; media references are links rather than downloaded files.

Inputs & outputs

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

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Extraction request`body`	`object`	Required	Required request object containing the page URLs and any chosen text format, focus or extraction settings. Required fields `urls`
Page URLs`body.urls`	`array`	Required	Provide 1–20 known page URLs. Each entry identifies a page to read; this list does not discover other pages on the site. Minimum items `1` Maximum items `20`
One page URL`body.urls[]`	`string`	—	One page address in the list, as a string of no more than 2,048 characters. Use the full address of the page whose text you need. Maximum length `2048`
Passage focus`body.query`	`string`	Optional	Optional keywords or a question to focus extraction on relevant highlights from the supplied pages. Omit it to request their main content instead.
Text format`body.format`	`string`	Optional	Choose markdown or text. The default is markdown; text requests plain text. This controls text presentation, not a copy of the visual page. Default `markdown` Options `markdown, text`
Time limit per page`body.timeout`	`integer`	Optional	Optional positive integer in seconds for each URL, default 30. It sets the extraction time limit rather than promising a completion time. Default `30` Maximum `9007199254740991`
Include audio references`body.include_audio`	`boolean`	Optional	Set true to request available audio URLs detected on the page. Default false. The references are not saved audio files or transcripts. Default `false`
Include image references`body.include_images`	`boolean`	Optional	Set true to request available image URLs from the page. Default false. Fetching or saving an image is a separate step. Default `false`
Include video references`body.include_videos`	`boolean`	Optional	Set true to request available video URLs detected on the page. Default false. This does not itself fetch video files or transcribe speech. Default `false`
Maximum cache age`body.max_age_seconds`	`integer`	Optional	Optional positive integer in seconds, default 86400 (one day). It sets how old cached content may be before a re-fetch, not the age of the article or a freshness guarantee. Default `86400` Maximum `9007199254740991`

View raw JSON

{
  "type": "object",
  "required": [
    "body"
  ],
  "properties": {
    "body": {
      "type": "object",
      "$schema": "https://json-schema.org/draft/2020-12/schema",
      "required": [
        "urls"
      ],
      "~standard": {
        "vendor": "zod",
        "version": 1,
        "jsonSchema": {}
      },
      "properties": {
        "urls": {
          "type": "array",
          "items": {
            "type": "string",
            "maxLength": 2048
          },
          "maxItems": 20,
          "minItems": 1,
          "description": "URLs to extract content from (max 20 per request). Failed URLs are returned with status 'failed' and are not billed."
        },
        "query": {
          "type": "string",
          "description": "Intent-focused keywords. When provided, returns query-relevant highlights per URL; otherwise the complete content."
        },
        "format": {
          "enum": [
            "markdown",
            "text"
          ],
          "type": "string",
          "default": "markdown",
          "description": "Format of the returned content."
        },
        "timeout": {
          "type": "integer",
          "default": 30,
          "maximum": 9007199254740991,
          "description": "Per-URL extraction timeout in seconds.",
          "exclusiveMinimum": 0
        },
        "include_audio": {
          "type": "boolean",
          "default": false
        },
        "include_images": {
          "type": "boolean",
          "default": false
        },
        "include_videos": {
          "type": "boolean",
          "default": false
        },
        "max_age_seconds": {
          "type": "integer",
          "default": 86400,
          "maximum": 9007199254740991,
          "description": "Maximum age (seconds) of cached content before a re-fetch.",
          "exclusiveMinimum": 0
        }
      },
      "additionalProperties": false
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

Read the extracted material in the context of its page URL. Main content is intended to remove page clutter, but it need not include every navigation item, table or visual element. Markdown and plain text are reading formats; they do not reproduce an interactive webpage or screenshot.

A focused query changes the reading task: use its returned highlights to find passages about that question. A passage is not a complete account of everything on the page. Your Agent can quote and compare the available text, but should not fill gaps with claims attributed to an unseen section. Octen’s Extract reference limits query to 500 characters.

Inspect each page separately. The URL, text and extraction status belong together. Titles, page-structure labels or category information can help when supplied; a label or success status alone does not prove that useful main content was available.

Keep dates in their original roles. A publication date describes the source material; a crawl date describes when it was obtained. The cache-age setting controls reuse before a re-fetch. Neither a recent crawl nor a smaller cache tolerance proves that the original claims are current. Octen’s Extract reference specifies 1–60 seconds for each URL’s timeout and 300–31,536,000 seconds for maximum cache age; it describes adjusting values outside those ranges to the nearest bound.

Requested media references may help you locate associated images, video or audio. They are available URLs, not fetched files or transcripts. Preparing structured notes, comparing sources or checking factual claims is separate Agent work based on material that has actually been read.

Octen: extracting text from supplied URLs

Octen: Extract request reference

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

It focuses extraction on relevant passages within the URLs you supply. It does not discover more URLs. Leave query out when you want the main page content, or use a focused question when you need passages about a particular issue.

Extract Text from Websites with Octen

Extract Text from Websites with Octen

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Choose the pages to read

Choose what part of each page to read

Check each page before comparing

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Choose the pages to read

Choose what part of each page to read

Check each page before comparing

Questions about this tool.

Does query search for new webpages?

Does a successful status mean I received the article?

How should I handle a failed URL in a batch?

Do the media options download files or create transcripts?

Will a smaller cache age prove that the page is current?

Will Markdown reproduce the original webpage layout?