> ## Documentation Index > Fetch the complete documentation index at: https://docs.olostep.com/llms.txt > Use this file to discover all available pages before exploring further. # Create Search > Search the web with a natural language query and get back a deduplicated list of relevant links with titles and descriptions. It will search for the query semantically across the web and return results. Optionally, pass `scrape_options` to also scrape every returned URL and embed `markdown_content` / `html_content` directly into each link. Page scrapes are billed automatically against your team via the underlying /v1/scrapes endpoint. See [Search feature](/features/search). ## OpenAPI ````yaml openapi/search.json POST /v1/searches openapi: 3.0.3 info: title: Search API version: 1.1.0 servers: - url: https://api.olostep.com security: [] paths: /v1/searches: post: summary: Create Search description: >- Search the web with a natural language query and get back a deduplicated list of relevant links with titles and descriptions. It will search for the query semantically across the web and return results. Optionally, pass `scrape_options` to also scrape every returned URL and embed `markdown_content` / `html_content` directly into each link. Page scrapes are billed automatically against your team via the underlying /v1/scrapes endpoint. See [Search feature](/features/search). requestBody: required: true content: application/json: schema: type: object properties: query: type: string description: The search query in natural language. example: What's going on with OpenAI's Sora shutting down? limit: type: integer description: Maximum number of links to return after deduplication. minimum: 1 maximum: 25 default: 12 example: 10 include_domains: type: array description: >- Restrict results to these domains. Bare hosts only — leading `http(s)://` and trailing slashes are stripped automatically. items: type: string example: - nytimes.com - wsj.com exclude_domains: type: array description: >- Exclude results from these domains. Bare hosts only — leading `http(s)://` and trailing slashes are stripped automatically. items: type: string example: - reddit.com scrape_options: $ref: '#/components/schemas/ScrapeOptions' required: - query examples: minimal: summary: Minimal — query only value: query: Best Answer Engine Optimization startups withScrape: summary: With scrape_options + domain filters + limit value: query: What's going on with OpenAI's Sora shutting down? limit: 10 include_domains: - nytimes.com - wsj.com exclude_domains: - pinterest.com scrape_options: formats: - markdown remove_css_selectors: default timeout: 25 responses: '200': description: Successful response with search results. content: application/json: schema: type: object properties: id: type: string example: search_9bi0sbj9xa object: type: string example: search created: type: integer example: 1760327323 metadata: type: object query: type: string example: Best Answer Engine Optimization startups credits_consumed: type: integer description: >- Total credits consumed by this request: 5 base search credits + sum of per-page scrape credits when scrape_options was used. example: 10 result: type: object properties: json_content: type: string description: Stringified JSON of the full search result. json_hosted_url: type: string description: URL to the hosted JSON file on S3. nullable: true links: type: array description: >- Deduplicated list of relevant links found across multiple searches. Each link may include `markdown_content` and/or `html_content` when scrape_options was provided. items: $ref: '#/components/schemas/SearchLink' size_exceeded: type: boolean description: >- True when inline per-link content exceeded the 9MB safety cap. When true, content fields are nulled in the response — use json_hosted_url to fetch the full payload. credits_consumed: type: integer description: Same value as the top-level credits_consumed. '400': description: >- Bad Request — validation failed (missing query, invalid limit, unsupported scrape_options.formats value, etc.). Response body is `{ "message": "" }`. '401': description: Invalid API Key '500': description: Internal Server Error security: - Authorization: [] components: schemas: ScrapeOptions: type: object description: >- Optional. When provided, each returned link is also scraped and its content embedded into the response. Per-page scrape credits are billed automatically by the underlying /v1/scrapes endpoint. properties: formats: type: array description: >- Output formats to request for each scraped page. Defaults to ['markdown'] when omitted. For now only 'html' and 'markdown' are supported on /v1/searches. items: type: string enum: - html - markdown default: - markdown example: - markdown remove_css_selectors: type: string description: >- Option to remove certain CSS selectors from the content. Forwarded to /v1/scrapes. Defaults to 'default' which strips ['nav','footer','script','style','noscript','svg',[role=alert],[role=banner],[role=dialog],[role=alertdialog],[role=region][aria-label*=skip i],[aria-modal=true]]. Pass 'none' to disable, or a JSON-stringified array of selectors to remove. default: default timeout: type: integer description: >- Wallclock budget in seconds for the entire scrape phase. After this elapses, the search returns with whatever links are available — content fields will be null for any links that hadn't finished scraping. minimum: 1 maximum: 60 default: 25 SearchLink: type: object properties: url: type: string description: The URL of the result. title: type: string nullable: true description: The title of the result page. description: type: string nullable: true description: A short snippet or description of the result. markdown_content: type: string nullable: true description: >- Markdown content of the page. Only present when scrape_options.formats includes 'markdown'. Null if the scrape failed, was empty, or hit the global timeout. For Reddit comment URLs, this is generated from the @olostep/reddit-post parser's structured JSON for higher quality. html_content: type: string nullable: true description: >- HTML content of the page. Only present when scrape_options.formats includes 'html'. Null if the scrape failed, was empty, or hit the global timeout. securitySchemes: Authorization: type: http scheme: bearer description: >- Bearer authentication header of the form Bearer , where is your auth token. ````