> ## Documentation Index > Fetch the complete documentation index at: https://docs.prefetch.io/llms.txt > Use this file to discover all available pages before exploring further. # GET /search > Search the web and get each result's title, URL and snippet — optionally with every result page scraped. 2 credits per 10 results. ## Overview `/search` runs a web search (Google results) and returns the results best first: each one's `url`, `title` and `description` (the snippet the search engine shows). Set `type=news` for recent articles with their publisher and date. Ask for page content and every result page is scraped in the same request, exactly as [`GET /scrape`](/api-reference/endpoint/scrape) would scrape it — markdown, a summary, links, structured JSON — and returned in `scrape` on its result. One call answers "find pages about X and give me their text". Use `/search` when you need to **find** pages. When you need a researched answer that cites its sources, use [`POST /answers`](/mcp/tools#prefetch_answer); when you already know the page, use `/scrape`. ## Example requests ```bash Results only theme={"theme":"css-variables"} curl "https://api.prefetch.io/search?query=open+source+CRM&limit=10" \ -H "X-API-Key: $PREFETCH_API_KEY" ``` ```bash With page content theme={"theme":"css-variables"} curl "https://api.prefetch.io/search?query=stripe+pricing&limit=5&formats=markdown" \ -H "X-API-Key: $PREFETCH_API_KEY" ``` ```bash POST, news from the past week theme={"theme":"css-variables"} curl -X POST "https://api.prefetch.io/search" \ -H "X-API-Key: $PREFETCH_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "query": "series A funding developer tools", "type": "news", "time_range": "week", "limit": 5, "scrape_options": { "formats": ["summary"] } }' ``` ```javascript Node.js theme={"theme":"css-variables"} const res = await fetch("https://api.prefetch.io/search", { method: "POST", headers: { "X-API-Key": process.env.PREFETCH_API_KEY, "Content-Type": "application/json" }, body: JSON.stringify({ query: "headless CMS comparison", limit: 5, exclude_domains: ["pinterest.com"], scrape_options: { formats: ["markdown"], only_main_content: true }, }), }); const { data } = await res.json(); for (const r of data.results) { console.log(r.position, r.url, r.scrape?.markdown?.content.slice(0, 200) ?? r.scrape_error?.error); } ``` ## Example response ```json theme={"theme":"css-variables"} { "success": true, "data": { "query": "stripe pricing", "type": "web", "results": [ { "position": 1, "url": "https://stripe.com/pricing", "title": "Pricing & Fees | Stripe", "description": "Find Stripe fees and pricing information. Find information on pricing for payments, billing…", "date": null, "source": null, "image_url": null, "scrape": { "url": "https://stripe.com/pricing", "final_url": "https://stripe.com/pricing", "domain": "stripe.com", "status_code": 200, "metadata": { "title": "Pricing & Fees | Stripe", "language": "en", "content_type": "text/html" }, "markdown": { "content": "# Pricing\n\nIntegrated per-transaction pricing…", "char_count": 18233, "word_count": 2611, "truncated": false, "main_content_only": true } } }, { "position": 2, "url": "https://www.g2.com/products/stripe/pricing", "title": "Stripe Pricing 2026", "description": "Stripe pricing and plans…", "date": null, "source": null, "image_url": null, "scrape_error": { "error": "The target URL could not be accessed. The site may be blocking automated requests.", "error_code": "TARGET_BLOCKED" } } ], "count": 2, "scraped": 1, "warnings": ["1 result page(s) could not be scraped; see each result's scrape_error. They were not charged."] }, "meta": { "requestId": "6f1d2c3b-4a5e-4f60-8b7a-9c0d1e2f3a4b", "durationMs": 8230 } } ``` ## Narrowing a search | Parameter | Does | | - | - | | `type` | `web` (default) or `news`. News results carry `source` (the publisher), `date` and `image_url`. | | `time_range` | `hour`, `day`, `week`, `month` or `year`. | | `country` / `language` | Two-letter codes, e.g. `country=de&language=de`. | | `location` | A city or region to search from, e.g. `Berlin, Germany`. | | `include_domains` | Search only these domains (up to 10). | | `exclude_domains` | Leave these out (up to 10). | Search operators work in `query` as they do on Google: `"exact phrase"`, `-word`, `filetype:pdf`, `site:example.com`. Results on your account's blocklist are dropped, and the rest renumbered. ## Scraping the results Ask for content with `formats` on GET (`formats=markdown,links`, plus `only_main_content` and `max_age`), or a `scrape_options` object on POST — every option [`/scrape`](/api-reference/endpoint/scrape) takes. Up to **10** results can be scraped in one request. Result pages are fetched in parallel within one deadline (45 s by default; `timeout` can shorten it). A page that cannot be fetched — blocked, too slow, not a page — gets a `scrape_error` with an [error code](/concepts/errors) instead of `scrape`, and is **not charged**. The search itself still succeeds. ## Credits | What | Credits | | - | - | | The search | **2 per 10 results** returned, rounded up — at least 2 | | Each result page scraped | 1, or 4 when `summary`, `json` or `check_prompt_injection` is asked for | | Result pages that could not be scraped | 0 | A 10-result search costs 2 credits; with every page as markdown, 12. ## Related One page, every format. **1 credit.** Hundreds of known URLs as one job. ## OpenAPI ````yaml GET /search openapi: 3.1.0 info: title: Prefetch API description: >- Extract brand identity, company data, screenshots, IAB classifications, and clean page content from any URL. version: 1.0.0 contact: email: support@prefetch.io servers: - url: https://api.prefetch.io description: Production security: - apiKey: [] tags: - name: Endpoints description: Data extraction endpoints. All require authentication and consume credits. - name: Health description: Health and readiness probes. No authentication required. paths: /search: get: tags: - Endpoints summary: Search the web description: >- Searches the web (Google results) and returns each result's title, URL and snippet, best first. `type=news` returns recent articles with their publisher and date. Narrow it with `time_range`, `country`, `language`, `location`, `include_domains` and `exclude_domains`. Ask for page content and each result page is scraped in the same request — through `formats` on GET, or a `scrape_options` object on POST, taking the options /scrape takes — and returned in `scrape` on its result. A page that cannot be fetched in time gets a `scrape_error` instead and is not charged. **Credit cost: 2 per 10 results** (rounded up), plus 1 per result page scraped (4 for `summary` or `json`). Failed requests are never charged. operationId: searchWeb parameters: - schema: type: string description: >- What to search for, as you would type it into Google (1–400 characters). Search operators work (`"exact phrase"`, `-word`, `filetype:pdf`). `q` is accepted too. example: best open source CRM required: true description: >- What to search for, as you would type it into Google (1–400 characters). Search operators work (`"exact phrase"`, `-word`, `filetype:pdf`). `q` is accepted too. name: query in: query - schema: type: string description: >- Results to return (1–50; 1–10 when results are scraped). Default: 10. example: '10' required: false description: >- Results to return (1–50; 1–10 when results are scraped). Default: 10. name: limit in: query - schema: type: string enum: - web - news description: >- `web` (default) or `news` — recent articles with their publisher and date. required: false description: >- `web` (default) or `news` — recent articles with their publisher and date. name: type in: query - schema: type: string description: Two-letter country code to search from, e.g. `us`, `de`. example: us required: false description: Two-letter country code to search from, e.g. `us`, `de`. name: country in: query - schema: type: string description: Two-letter language code for the results, e.g. `en`, `de`. example: en required: false description: Two-letter language code for the results, e.g. `en`, `de`. name: language in: query - schema: type: string description: A city or region to search from, e.g. `Berlin, Germany`. required: false description: A city or region to search from, e.g. `Berlin, Germany`. name: location in: query - schema: type: string enum: - hour - day - week - month - year description: Only results from the past hour, day, week, month or year. required: false description: Only results from the past hour, day, week, month or year. name: time_range in: query - schema: type: string description: Comma-separated domains to search within (up to 10). example: github.com,gitlab.com required: false description: Comma-separated domains to search within (up to 10). name: include_domains in: query - schema: type: string description: Comma-separated domains to leave out (up to 10). example: pinterest.com required: false description: Comma-separated domains to leave out (up to 10). name: exclude_domains in: query - schema: type: string description: >- Scrape each result page into these formats (the /scrape formats: markdown, summary, html, raw_html, links, images, json, screenshot). Omit to return search results only. example: markdown required: false description: >- Scrape each result page into these formats (the /scrape formats: markdown, summary, html, raw_html, links, images, json, screenshot). Omit to return search results only. name: formats in: query - schema: type: string description: >- With `formats`: strip navigation, sidebars and footers from each page. Default: true. required: false description: >- With `formats`: strip navigation, sidebars and footers from each page. Default: true. name: only_main_content in: query - schema: type: string description: >- With `formats`: accept a stored copy of a page fetched at most this many milliseconds ago. example: '86400000' required: false description: >- With `formats`: accept a stored copy of a page fetched at most this many milliseconds ago. name: max_age in: query - schema: type: string description: >- Give up after this many milliseconds (5,000–45000). Result pages still loading are returned with a `scrape_error` and not charged. example: '30000' required: false description: >- Give up after this many milliseconds (5,000–45000). Result pages still loading are returned with a `scrape_error` and not charged. name: timeout in: query responses: '200': description: Results, best first. content: application/json: schema: type: object properties: success: type: boolean enum: - true data: $ref: '#/components/schemas/SearchData' meta: $ref: '#/components/schemas/Meta' required: - success - data - meta example: success: true data: query: stripe pricing type: web count: 2 scraped: 0 warnings: [] results: - position: 1 url: https://stripe.com/pricing title: Pricing & Fees | Stripe description: Find Stripe fees and pricing information… date: null source: null image_url: null - position: 2 url: https://docs.stripe.com/billing/subscriptions/pricing title: Pricing models | Stripe Documentation description: Learn about common pricing models… date: null source: null image_url: null meta: requestId: 6f1d2c3b-4a5e-4f60-8b7a-9c0d1e2f3a4b durationMs: 1840 '400': description: Invalid or missing parameters. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '401': description: Missing or invalid API key. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '403': description: Revoked key, expired key, blocked URL, or credit limit. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '429': description: Rate or concurrency limit exceeded. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '503': description: Service temporarily unavailable — retry shortly. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' components: schemas: SearchData: type: object properties: query: type: string type: type: string enum: - web - news results: type: array items: $ref: '#/components/schemas/SearchResultItem' count: type: number scraped: type: number description: Result pages fetched — each charged at the /scrape rate. warnings: type: array items: type: string required: - query - type - results - count - scraped - warnings Meta: type: object properties: requestId: type: string format: uuid description: Unique identifier for this request. example: a3f2c1d4-7b6e-4f2a-9c1d-8e3f2a1b4c5d durationMs: type: integer nullable: true description: Time taken to process the request in milliseconds. example: 1842 required: - requestId - durationMs ErrorResponse: type: object properties: success: type: boolean enum: - false description: >- Always false. Absent, like `meta`, on authentication and rate-limit errors, which are answered before a request is processed. error: type: string description: >- Human-readable error message. The wording can change: branch on `error_code`. example: Credit limit exceeded error_code: type: string enum: - INVALID_REQUEST - PAYLOAD_TOO_LARGE - NOT_FOUND - CONFLICT - UNAUTHORIZED - FORBIDDEN - CREDITS_EXHAUSTED - RATE_LIMITED - CONCURRENCY_LIMITED - URL_BLOCKED - URL_NOT_PERMITTED - DNS_RESOLUTION_FAILED - TARGET_BLOCKED - TARGET_CONNECTION_FAILED - TARGET_TIMEOUT - TARGET_SSL_ERROR - TARGET_SERVER_ERROR - TARGET_CLIENT_ERROR - TARGET_NOT_A_PAGE - TARGET_UNLOADABLE - RESEARCH_INCONCLUSIVE - TIMEOUT - SERVICE_BUSY - SERVICE_UNAVAILABLE - INTERNAL_ERROR description: >- What went wrong, for machines. Stable: a code keeps its meaning once published. See [Error handling](/concepts/errors). example: CREDITS_EXHAUSTED retryable: type: boolean description: >- Whether the same request can succeed if sent again after a short wait (honour `Retry-After` when present). False for problems with the request or the target. example: false retryAfterSeconds: type: integer description: 'On `RATE_LIMITED`: seconds until a request will be accepted.' limit: type: integer description: >- On `CONCURRENCY_LIMITED`: how many requests this key may have in flight. meta: $ref: '#/components/schemas/Meta' required: - error - error_code - retryable SearchResultItem: type: object properties: position: type: number description: Rank on the results page, from 1. example: 1 url: type: string title: type: string nullable: true description: type: string nullable: true description: The snippet the search engine shows for the result. date: type: string nullable: true description: Publication date as the result shows it, when it shows one. example: 2 days ago source: type: string nullable: true description: 'News results: the publisher.' example: The Verge image_url: type: string nullable: true description: 'News results: the article''s thumbnail.' scrape: allOf: - $ref: '#/components/schemas/ScrapeData' - description: >- The result page in the requested formats — the same object GET /scrape returns. Present when scraping was asked for and the page was fetched. scrape_error: type: object properties: error: type: string error_code: type: string enum: - INVALID_REQUEST - PAYLOAD_TOO_LARGE - NOT_FOUND - CONFLICT - UNAUTHORIZED - FORBIDDEN - CREDITS_EXHAUSTED - RATE_LIMITED - CONCURRENCY_LIMITED - URL_BLOCKED - URL_NOT_PERMITTED - DNS_RESOLUTION_FAILED - TARGET_BLOCKED - TARGET_CONNECTION_FAILED - TARGET_TIMEOUT - TARGET_SSL_ERROR - TARGET_SERVER_ERROR - TARGET_CLIENT_ERROR - TARGET_NOT_A_PAGE - TARGET_UNLOADABLE - TARGET_ELEMENT_NOT_FOUND - RESEARCH_INCONCLUSIVE - TIMEOUT - SERVICE_BUSY - SERVICE_UNAVAILABLE - INTERNAL_ERROR required: - error - error_code description: Why the result page could not be scraped. Not charged. required: - position - url - title - description - date - source - image_url ScrapeData: allOf: - $ref: '#/components/schemas/ScrapePayload' required: - url - final_url - metadata ScrapePayload: type: object properties: url: type: string final_url: type: string domain: type: string nullable: true example: stripe.com status_code: type: integer nullable: true metadata: $ref: '#/components/schemas/ScrapeMetadata' markdown: $ref: '#/components/schemas/MarkdownBlock' summary: $ref: '#/components/schemas/SummaryBlock' html: $ref: '#/components/schemas/HtmlBlock' raw_html: $ref: '#/components/schemas/HtmlBlock' links: $ref: '#/components/schemas/LinksBlock' images: $ref: '#/components/schemas/ScrapeImagesBlock' json: $ref: '#/components/schemas/JsonBlock' screenshot: $ref: '#/components/schemas/ScrapeScreenshotBlock' pdf: $ref: '#/components/schemas/PdfBlock' prompt_injection: $ref: '#/components/schemas/PromptInjectionBlock' warnings: type: array items: type: string description: Non-fatal problems, such as a format that could not be produced. description: >- The page content fields. Which of them are present depends on the `formats` requested. ScrapeMetadata: type: object description: Always present, whatever formats were requested. properties: source_url: type: string nullable: true description: The URL you requested. final_url: type: string nullable: true description: The URL after redirects. status_code: type: integer nullable: true title: type: string nullable: true example: Stripe | Financial Infrastructure description: type: string nullable: true language: type: string nullable: true example: en canonical_url: type: string nullable: true favicon: type: string nullable: true author: type: string nullable: true site_name: type: string nullable: true published_time: type: string nullable: true modified_time: type: string nullable: true robots: type: string nullable: true description: The page's own robots meta tag. keywords: type: array items: type: string open_graph: type: object additionalProperties: type: string twitter: type: object additionalProperties: type: string json_ld: type: array items: type: object description: Parsed JSON-LD blocks, up to 10. headings: type: array description: The page's heading outline, up to 100 entries. items: type: object properties: level: type: integer text: type: string content_type: type: string enum: - text/html - application/pdf description: >- What the URL served. A PDF's text becomes the markdown; it has no html. page_count: type: integer description: 'PDFs only: pages in the document.' cache_state: type: string enum: - hit - miss description: >- Present when the request set `max_age`: `hit` = served from a stored fetch, `miss` = fetched live. cached_at: type: string description: 'On a cache hit: when the stored copy was fetched (ISO 8601).' MarkdownBlock: type: object properties: content: type: string description: >- GitHub Flavored Markdown. Links and images are resolved to absolute URLs. char_count: type: integer word_count: type: integer truncated: type: boolean description: True when the content hit the 1 MB size cap and was cut. main_content_only: type: boolean description: >- True when boilerplate was successfully isolated away. False means the full page body was returned. SummaryBlock: type: object properties: text: type: string nullable: true description: Null when the summary could not be produced; see `warnings`. model: type: string nullable: true HtmlBlock: type: object properties: content: type: string char_count: type: integer truncated: type: boolean main_content_only: type: boolean LinksBlock: type: object properties: internal: type: array description: Links on the same site, subdomains included. items: $ref: '#/components/schemas/ScrapeLinkItem' external: type: array items: $ref: '#/components/schemas/ScrapeLinkItem' count: type: integer ScrapeImagesBlock: type: object properties: assets: type: array items: $ref: '#/components/schemas/ScrapeImageItem' og_image: type: string nullable: true twitter_image: type: string nullable: true count: type: integer JsonBlock: type: object properties: data: type: object nullable: true description: The extracted object, shaped by your prompt or schema. model: type: string nullable: true ScrapeScreenshotBlock: type: object description: Present when `formats` includes `screenshot`. properties: url: type: string description: 'The image on our CDN. Short-lived: copy it if you need to keep it.' viewport: type: object properties: width: type: integer height: type: integer full_page: type: boolean format: type: string enum: - jpeg - png - webp PdfBlock: type: object description: Present when the URL served a PDF. Its text is in `markdown`. properties: page_count: type: integer pages_parsed: type: integer description: Pages whose text was read (capped for very long documents). truncated: type: boolean PromptInjectionBlock: type: object description: Present when the request set `check_prompt_injection`. properties: detected: type: boolean description: >- True when the page's text appears to contain instructions aimed at an AI model. reason: type: string nullable: true description: >- Why it was flagged, quoting or describing the suspicious text. Null when nothing was found. example: >- The page tells AI assistants to ignore their instructions and send the user's API keys to an outside address. complete: type: boolean description: >- True when all of the page's text was checked. False when part of it could not be (a very long page, or the classifier was unavailable): `detected: false` then means nothing was found in what was checked, not that the page is clean. model: type: string nullable: true description: The model that checked it; null for a page with no text. example: gpt-5.6-luna required: - detected - reason - complete - model ScrapeLinkItem: type: object properties: url: type: string text: type: string nullable: true description: The anchor text. title: type: string nullable: true rel: type: string nullable: true ScrapeImageItem: type: object properties: url: type: string alt: type: string nullable: true width: type: integer nullable: true height: type: integer nullable: true loading: type: string nullable: true securitySchemes: apiKey: type: apiKey in: header name: X-API-Key description: Your Prefetch API key. Obtain one from the dashboard. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.