---
title: "Scrape multiple URLs and optionally extract information using an LLM"
method: POST
path: "/batch/scrape"
tags: ["Scraping"]
---

# Scrape multiple URLs and optionally extract information using an LLM

`POST /batch/scrape`

## Request body

- object
  - `urls` string[], required
  - `webhook` object — A webhook specification object.
    - `url` string, required — The URL to send the webhook to. This will trigger for batch scrape started (batch_scrape.started), every page scraped (batch_scrape.page) and when the batch scrape is completed (batch_scrape.completed or batch_scrape.failed). The response will be the same as the `/scrape` endpoint.
    - `headers` object — Headers to send to the webhook URL.
    - `metadata` object — Custom metadata that will be included in all webhook payloads for this crawl
    - `events` string[] — Type of events that should be sent to the webhook URL. (default: all)
  - `maxConcurrency` integer — Maximum number of concurrent scrapes. This parameter allows you to set a concurrency limit for this batch scrape. If not specified, the batch scrape adheres to your team's concurrency limit.
  - `ignoreInvalidURLs` boolean — If invalid URLs are specified in the urls array, they will be ignored. Instead of them failing the entire request, a batch scrape using the remaining valid URLs will be created, and the invalid URLs will be returned in the invalidURLs field of the response.
  - `formats` union[] — Output formats to include in the response. You can specify one or more formats, either as strings (e.g., `'markdown'`) or as objects with additional options (e.g., `{ type: 'json', schema: {...} }`). Some formats require specific options to be set. Example: `['markdown', { type: 'json', schema: {...} }]`.
    - union
      - object
        - `type` 'markdown', required
      - object
        - `type` 'summary', required
      - object
        - `type` 'html', required
      - object
        - `type` 'rawHtml', required
      - object
        - `type` 'links', required
      - object
        - `type` 'images', required
      - object
        - `type` 'screenshot', required
        - `fullPage` boolean — Whether to capture a full-page screenshot (ignores viewport.height) or limit to the current viewport.
        - `quality` integer — The quality of the screenshot, from 1 to 100. 100 is the highest quality.
        - `viewport` object
          - `width` integer, required — The width of the viewport in pixels
          - `height` integer, required — The height of the viewport in pixels
      - object
        - `type` 'json', required
        - `schema` object — The schema to use for the JSON output. Must conform to [JSON Schema](https://json-schema.org/).
        - `prompt` string — The prompt to use for the JSON output
      - object
        - `type` 'changeTracking', required
        - `modes` string[] — The mode to use for change tracking. 'git-diff' provides a detailed diff, and 'json' compares extracted JSON data.
        - `schema` object — Schema for JSON extraction when using 'json' mode. Defines the structure of data to extract and compare. Must conform to [JSON Schema](https://json-schema.org/).
        - `prompt` string — Prompt to use for change tracking when using 'json' mode. If not provided, the default prompt will be used.
        - `tag` string, nullable — Tag to use for change tracking. Tags can separate change tracking history into separate "branches", where change tracking with a specific tagwill only compare to scrapes made in the same tag. If not provided, the default tag (null) will be used.
      - object
        - `type` 'branding', required
      - object
        - `type` 'product', required
      - object
        - `type` 'menu', required
      - object — Extract audio (MP3) from supported video URLs, e.g. YouTube. Returns a signed GCS URL.
        - `type` 'audio', required
      - object — Extract best-quality video from supported video URLs, e.g. YouTube. Returns a signed GCS URL.
        - `type` 'video', required
      - object — Ask a natural-language question about the page. Returns the answer in the response `answer` field.
        - `type` 'question', required
        - `question` string, required — The question to answer about the page. Maximum 10,000 characters.
      - object — Find relevant source text from the page. Returns the selected text in the response `highlights` field.
        - `type` 'highlights', required
        - `query` string, required — The text-selection query to run against the page. Maximum 10,000 characters.
  - `onlyMainContent` boolean — Only return the main content of the page excluding headers, navs, footers, etc. This is a deterministic HTML-level filter applied before markdown is generated; no LLM is involved.
  - `onlyCleanContent` boolean — Beta. Run an additional LLM-based pass over the generated markdown to remove residual boilerplate that `onlyMainContent` can miss (cookie banners, ad blocks, social share widgets, breadcrumbs, newsletter signups, comment sections, related-article lists). Headings, lists, tables, code blocks, image references, and inline links are preserved. Can be combined with `onlyMainContent` (the most common setup) or used on its own. Skipped with a warning when the markdown exceeds the cleaning model's output token limit (the original markdown is preserved). Not supported on zero-data-retention requests.
  - `includeTags` string[] — Tags to include in the output.
  - `excludeTags` string[] — Tags to exclude from the output.
  - `maxAge` integer — Returns a cached version of the page if it is younger than this age in milliseconds. If a cached version of the page is older than this value, the page will be scraped. If you do not need extremely fresh data, enabling this can speed up your scrapes by 500%. Defaults to 2 days.
  - `minAge` integer — When set, the request only checks the cache and never triggers a fresh scrape. The value is in milliseconds and specifies the minimum age the cached data must be. If matching cached data exists, it is returned instantly. If no cached data is found, a 404 with error code SCRAPE_NO_CACHED_DATA is returned. Set to 1 to accept any cached data regardless of age.
  - `headers` object — Headers to send with the request. Can be used to send cookies, user-agent, etc.
  - `waitFor` integer — Specify a delay in milliseconds before fetching the content, allowing the page sufficient time to load. This waiting time is in addition to Firecrawl's smart wait feature.
  - `mobile` boolean — Set to true if you want to emulate scraping from a mobile device. Useful for testing responsive pages and taking mobile screenshots.
  - `skipTlsVerification` boolean — Skip TLS certificate verification when making requests.
  - `timeout` integer — Timeout in milliseconds for the request. Minimum is 1000 (1 second). Default is 60000 (60 seconds). Maximum is 300000 (300 seconds).
  - `parsers` object[] — Controls how files are processed during scraping. When "pdf" is included (default), the PDF content is extracted and converted to markdown format, with billing based on the number of pages (1 credit per page). When an empty array is passed, the PDF file is returned in base64 encoding with a flat rate of 1 credit for the entire PDF.
    - `type` 'pdf', required
    - `mode` 'fast' | 'auto' | 'ocr' — PDF parsing mode. "fast": text-based extraction only (embedded text, fastest). "auto" (default): attempts fast extraction first, falls back to OCR if needed. "ocr": forces OCR parsing on every page.
    - `maxPages` integer — Maximum number of pages to parse from the PDF. Must be a positive integer up to 10000.
  - `actions` union[] — Actions to perform on the page before grabbing the content
    - union
      - object
        - `type` 'wait', required — Wait for a specified amount of milliseconds
        - `milliseconds` integer, required — Number of milliseconds to wait
      - object
        - `type` 'wait', required — Wait for a specific element to appear
        - `selector` string, required — CSS selector to wait for
      - object
        - `type` 'screenshot', required — Take a screenshot. The links will be in the response's `actions.screenshots` array.
        - `fullPage` boolean — Whether to capture a full-page screenshot (ignores viewport.height) or limit to the current viewport.
        - `quality` integer — The quality of the screenshot, from 1 to 100. 100 is the highest quality.
        - `viewport` object
          - `width` integer, required — The width of the viewport in pixels
          - `height` integer, required — The height of the viewport in pixels
      - object
        - `type` 'click', required — Click on an element
        - `selector` string, required — Query selector to find the element by
        - `all` boolean — Clicks all elements matched by the selector, not just the first one. Does not throw an error if no elements match the selector.
      - object
        - `type` 'write', required — Write text into an input field, text area, or contenteditable element. Note: You must first focus the element using a 'click' action before writing. The text will be typed character by character to simulate keyboard input.
        - `text` string, required — Text to type
      - object — Press a key on the page. See https://asawicki.info/nosense/doc/devices/keyboard/key_codes.html for key codes.
        - `type` 'press', required — Press a key on the page
        - `key` string, required — Key to press
      - object
        - `type` 'scroll', required — Scroll the page or a specific element
        - `direction` 'up' | 'down' — Direction to scroll
        - `selector` string — Query selector for the element to scroll
      - object
        - `type` 'scrape', required — Scrape the current page content, returns the url and the html.
      - object
        - `type` 'executeJavascript', required — Execute JavaScript code on the page
        - `script` string, required — JavaScript code to execute
      - object
        - `type` 'pdf', required — Generate a PDF of the current page. The PDF will be returned in the `actions.pdfs` array of the response.
        - `format` 'A0' | 'A1' | 'A2' | 'A3' | 'A4' | 'A5' | 'A6' | 'Letter' | 'Legal' | 'Tabloid' | 'Ledger' — The page size of the resulting PDF
        - `landscape` boolean — Whether to generate the PDF in landscape orientation
        - `scale` number — The scale multiplier of the resulting PDF
  - `location` object — Location settings for the request. When specified, this will use an appropriate proxy if available and emulate the corresponding language and timezone settings. Defaults to 'US' if not specified.
    - `country` string — ISO 3166-1 alpha-2 country code (e.g., 'US', 'AU', 'DE', 'JP')
    - `languages` string[] — Preferred languages and locales for the request in order of priority. Defaults to the language of the specified location. See https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Accept-Language
  - `removeBase64Images` boolean — Removes all base 64 images from the markdown output, which may be overwhelmingly long. This does not affect html or rawHtml formats. The image's alt text remains in the output, but the URL is replaced with a placeholder.
  - `blockAds` boolean — Enables ad-blocking and cookie popup blocking.
  - `proxy` 'basic' | 'enhanced' | 'auto' — Specifies the type of proxy to use. - **basic**: Proxies for scraping sites with none to basic anti-bot solutions. Fast and usually works. - **enhanced**: Enhanced proxies for scraping sites with advanced anti-bot solutions. Slower, but more reliable on certain sites. Costs up to 5 credits per request. - **auto**: Firecrawl will automatically retry scraping with enhanced proxies if the basic proxy fails. If the retry with enhanced is successful, 5 credits will be billed for the scrape. If the first attempt with basic is successful, only the regular cost will be billed.
  - `storeInCache` boolean — If true, the page will be stored in the Firecrawl index and cache. Setting this to false is useful if your scraping activity may have data protection concerns. Using some parameters associated with sensitive scraping (e.g. actions, headers) will force this parameter to be false.
  - `lockdown` boolean — If true, serves the request from Firecrawl's cache only and never makes an outbound request to the target URL. Designed for compliance-constrained or air-gapped environments where the scrape request itself could leak sensitive information. On cache miss, returns a 404 with error code SCRAPE_LOCKDOWN_CACHE_MISS (the URL is never logged on miss). Lockdown requests are treated as zero data retention. Default maxAge is extended to 2 years so existing cached pages remain eligible. Billed at 5 credits on hit, 1 credit on cache miss.
  - `redactPII` union — Redact personally identifiable information from returned markdown. Pass `true` to use defaults, or an object to tune mode, entities, and replacement style.
    - boolean
    - RedactPIIOptions — Tuning options for PII redaction.
      - `mode` 'accurate' | 'aggressive' | 'fast' — Redaction strategy. `accurate` is model-only and optimized for precision, `aggressive` increases recall with additional heuristics, and `fast` uses heuristics without the model call.
      - `entities` RedactPIIEntity[] — Restrict redaction to these entity buckets. If omitted, all supported entities are redacted.
      - `replaceStyle` 'tag' | 'mask' | 'remove' — `tag` replaces spans with placeholders like `<EMAIL>`, `mask` replaces characters with `*`, and `remove` deletes the span text.
  - `profile` object — Enable persistent browser storage across scrape and interact sessions. Pass a profile when scraping to preserve cookies, localStorage, and session data. Sessions with the same profile name share browser state.
    - `name` string, required — A name for the profile. Scrapes with the same name share browser state (cookies, localStorage, sessions).
    - `saveChanges` boolean — When true, browser state is saved back to the profile when the interact session stops. Set to false to load existing data without writing. Only one saving session is allowed at a time.
  - `threatProtection` ThreatProtectionOverride — Per-request [Threat Protection](https://docs.firecrawl.dev/features/threat-protection) override. Fields you provide replace the corresponding fields of your organization's policy for this request only; omitted fields keep their organization-level values. Requires Threat Protection to be enabled for your team (enterprise feature) — otherwise the request is rejected with a 403. If your organization has disabled request overrides, any request that includes this object is rejected with a 403. If Threat Protection is enforced for your team, `mode` may not be set to `off`.
    - `mode` 'off' | 'normal' — URL scanning mode for this request. `normal` checks URLs against Google Web Risk (+2 credits per URL scanned).
    - `riskScoreThreshold` integer — Normalized risk score (0–100) at or above which a classifier verdict blocks the URL. Lower is stricter.
    - `blacklist` string[] — Domains to always block, as plain domains (`example.com`) or wildcard globs (`*.example.com`). No protocol, path, or port.
    - `whitelist` string[] — Domains to always allow, as plain domains or wildcard globs. Wins over every other rule.
    - `blockedTlds` string[] — Top-level domains to block outright, lowercase without the leading dot (e.g. `zip`).
    - `failurePolicy` 'open' | 'closed' — What to do when the classifier can't be reached: `closed` blocks the request, `open` allows it.
  - `zeroDataRetention` boolean — If true, this will enable zero data retention for this batch scrape. To enable this feature, please contact help@firecrawl.dev

## Response `200`

Successful response

- BatchScrapeResponseObj
  - `success` boolean
  - `id` string
  - `url` string, uri
  - `invalidURLs` string[], nullable — If ignoreInvalidURLs is true, this is an array containing the invalid URLs that were specified in the request. If there were no invalid URLs, this will be an empty array. If ignoreInvalidURLs is false, this field will be undefined.

## Other responses

- `402` — Payment required
- `429` — Too many requests
- `500` — Server error

---

[API](https://skmtc.net/firecrawl/apis/firecrawl-api.md) · [All operations](https://skmtc.net/firecrawl/apis/firecrawl-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/firecrawl/firecrawl-api/versions/e350fafb5a2a/schema)