> ## Documentation Index > Fetch the complete documentation index at: https://docs.langsearch.com/llms.txt > Use this file to discover all available pages before exploring further. # LangSearch Web Search API > Search the web, retrieve useful context, and build with source-backed information.
FREE WEB SEARCH API

The World Engine
for AGI.

Bring the world's information into your AI workflows. Search the web, retrieve useful context, and keep the sources.

Make your first search →Building with an agent ↗
LangSearch connects AI applications to information on the web. Send a query and receive structured results with source URLs and search snippets, or full webpage text when requested. Building with a coding agent? Share the [self-contained integration guide](/reference/search-api-guide-for-coding-agents). ## What you can build Give your model search context and source URLs to support its answers. Gather evidence across sources and refine the next search as you learn. Find relevant documentation and technical explanations for a development task. ## Make your first search ### 1. Get an API key Open the [Dashboard](https://langsearch.com/dashboard), choose **API keys**, and create a key. Replace `YOUR_LANGSEARCH_API_KEY` below with your key. Run the request from a server or trusted local environment. ### 2. Send a request Choose cURL, Python, or JavaScript. These examples use standard HTTP clients; no LangSearch-specific SDK is required. ```bash cURL theme={null} curl https://api.langsearch.com/v1/web-search \ -H "Authorization: Bearer YOUR_LANGSEARCH_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "query": "How does a search API help an AI agent?", "count": 5, "contents": {"text": true}, "freshness": "noLimit" }' ``` ```python Python theme={null} import json from urllib.request import Request, urlopen request = Request( "https://api.langsearch.com/v1/web-search", data=json.dumps({ "query": "How does a search API help an AI agent?", "count": 5, "contents": {"text": True}, "freshness": "noLimit" }).encode(), headers={ "Authorization": "Bearer YOUR_LANGSEARCH_API_KEY", "Content-Type": "application/json" } ) with urlopen(request, timeout=30) as response: result = json.load(response) for source in result["data"]["webPages"]["value"]: print(source["name"], source["url"]) ``` ```javascript JavaScript theme={null} const response = await fetch("https://api.langsearch.com/v1/web-search", { method: "POST", headers: { Authorization: "Bearer YOUR_LANGSEARCH_API_KEY", "Content-Type": "application/json", }, body: JSON.stringify({ query: "How does a search API help an AI agent?", count: 5, contents: { text: true }, freshness: "noLimit", }), signal: AbortSignal.timeout(30000), }); if (!response.ok) throw new Error(`HTTP ${response.status}`); const result = await response.json(); for (const source of result.data.webPages.value) { console.log(source.name, source.url); } ``` ### 3. Read the sources Results are in `data.webPages.value`. Each item represents a source: | Field | What it gives you | | --------------- | -------------------------------------------------- | | `name` | Source title | | `url` | Source URL to retain for citations | | `snippet` | Search text when full webpage text is not enabled | | `text` | Full webpage text when enabled; replaces `snippet` | | `datePublished` | Publication metadata, when available | The examples above enable `contents.text`, so read `text` for each source. Omit this option or set it to `false` to use `snippet` instead. Your model can use source text to compose an answer with citations. Text and date fields may be missing; an empty result array is also a valid response. ### Result metadata and missing fields Preserve result order unless your application has a reason to reorder it. For context, use `text` in text mode or `snippet` in snippet mode, and keep the source URL even if text is missing. Broaden the query or freshness window when the evidence is insufficient. ## Authentication and API keys Every request uses the same authentication header: ```http theme={null} Authorization: Bearer YOUR_LANGSEARCH_API_KEY ``` Create a key in **Dashboard → API keys**, copy it when the full key is shown, and store it securely. Run requests from a server or trusted local tool; do not expose the key in a browser bundle, public repository, or shared configuration file. To rotate a key, create a replacement, update your application, verify a request, and then delete the old key. All keys on your account share the same allowance. If you receive `401`, check the `Bearer ` prefix, accidental whitespace, an unchanged placeholder, or a deleted key. MCP uses the same Bearer authentication with its [client-specific configuration](/integrations/mcp). ## Choose what to retrieve | Your task | Starting point | | -------------------------- | ---------------------------------------------------- | | A focused factual question | `count: 5`, `contents: { "text": true }` | | A source list | `count: 10`; omit `contents` | | Recent developments | `contents: { "text": true }`, `freshness: "oneWeek"` | | A bounded model context | `contents: { "text": { "max_characters": 3000 } }` | `count` defaults to 10 and has a maximum of 50. `contents.text: true` requests full webpage text with a **5000-character limit per result**. An object also enables text: `contents: { "text": { "max_characters": 3000 } }`. The limit must be a positive integer, and you do not need a separate `true` flag. The limit is an upper bound; available text may be shorter. ## Filter by date and domain `freshness` accepts `noLimit` (default), `oneDay`, `oneWeek`, `oneMonth`, and `oneYear`, plus: * A specific UTC date: `2026-09-12`. * An inclusive UTC date range: `2026-09-01..2026-09-13`. Use valid calendar dates with the start on or before the end. Use `includeDomains` to restrict sources and `excludeDomains` to leave sources out. Domain entries are strings such as `langsearch.com` or `openai.com`; omit the arrays or leave them empty to apply no domain filter. ```json theme={null} { "query": "AI search APIs", "count": 10, "freshness": "2026-09-01..2026-09-13", "includeDomains": ["langsearch.com", "openai.com"], "excludeDomains": ["reddit.com"], "contents": { "text": { "max_characters": 3000 } } } ``` Date filtering uses source metadata, not a guarantee that a page was crawled during that window. For query refinement and choosing the right amount of context, read [Best Practices](/reference/search-best-practices). ## Connect your tools Connect Codex, Claude Code, Cursor, VS Code, and other clients to the official hosted server. Give your agent the official installation instructions. ## Next steps * [Web Search API reference](/api/web-search-api): request fields and interactive playground. * [Plan & usage](/limits/api-limits): account allowance and automatic resets. * [Errors & troubleshooting](/api/errors): diagnose failures and decide when to retry.