Categories · Data & Web

Web Search APIs

Every web search apis we track, compared.

TL;DR: Top pick: Firecrawl. 12 APIs compared, 5 with a free tier; cheapest published entry $0.001 API call (Linkup). Why →

Looking for a recommendation? See the Best Web Search APIs guide. · Last verified 2026-06-14

What is a Web Search API?

A Web Search API lets developers integrate web search apis capabilities over HTTP. The Web Search APIs below are compared on pricing, compliance, capabilities, and developer experience from structured, cited data.

To be listed as a Web Search API, an API must have a public HTTP endpoint and published, sourced data. Listings are ranked on those verifiable fields, never on payment.

By job: AI Answer-Engine APIs · Web Search APIs for RAG & LLM Grounding

Sorted by a data-readiness score (published pricing, free tier, self-serve access, compliance, webhooks/sandbox, capability breadth). No paid placement. How we rank. yes · no · · unknown.

12 APIs compared
#APIPricingFreeSelf-serveSOC 2HIPAAGDPRWebhooksActions
1FirecrawlSubscriptionType II·10
2ExaUsageType II·13
3Brave Search APIUsageType II12
4Tavily Search APIHybridType II7
5LinkupUsageType II5
6Jina AI Search Foundation (DeepSearch / Reader)Usage··17
7Parallel Search APIUsageType II··8
8You.com Search APIUsageType II9
9Kagi Search APIUsage - 10
10Perplexity Sonar APIUsageType II·9
11Mojeek Web Search APIUsage··9
12Google Programmable Search Engine (Custom Search JSON API)Usage···3

The APIs

  • #1 Firecrawl

    78 / 100

    Firecrawl is a REST API to search, scrape, and crawl the web at scale, built for AI agents, RAG pipelines, deep research, and lead enrichment, turning sites into LLM-ready data. It uses API-key auth, webhooks, six SDKs, and an official MCP server. Pricing is published and self-serve: 1,000 credits/month free, with subscriptions from $16/month. It carries SOC 2 Type 2, GDPR, and a published SLA. Used by Shopify, Canva, and Zapier.

    PricingSubscription · from $16 month · free tier
    TrustSOC 2 Type II · GDPR
    Used byShopify, Lovable, Canva, Zapier

    Firecrawl profile →

  • #2 Exa

    74 / 100

    Exa is a REST API for web search, content crawling, and research agents, designed specifically as a grounding and retrieval layer for LLM applications and AI agents. It offers neural search, full-page content extraction, news and research paper search, scheduled monitors, and an answer-with-citations endpoint, with SDKs for Python, JavaScript, and an MCP server. Pricing is usage-based at $1.00 per 1,000 requests, with a free tier of 20,000 requests per month and self-serve signup; an enterprise plan is available. The service holds SOC 2 Type 2 and HIPAA certifications, and customers include Cursor, Notion AI, HubSpot, and Databricks.

    PricingUsage · from $1 1,000 requests · free tier
    TrustSOC 2 Type II · HIPAA
    Does
    • AI answer
    • Semantic / neural
    • Full content for RAG
    • News search
    • Agent-ready / MCP
    • Zero data retention
    Used byCursor, Cognition, HubSpot, OpenRouter

    Exa profile →

  • #3 Brave Search API

    59 / 100

    Brave Search API gives developers access to Brave's independent web index of over 30 billion pages for building AI products, RAG pipelines, chatbots, and search engines, with endpoints covering web, news, image, video, local place search, answers, and LLM context. Pricing is usage-based at $4 per 1,000 requests, with a free tier of roughly $5 in monthly credits, self-serve signup, and an enterprise plan available. The API is SOC 2 Type 2 certified and GDPR compliant, ships with an MCP server, and counts Cohere, Mistral AI, AWS, Shopify, and Snowflake among its documented customers. Usage terms prohibit storing results or using them to train AI models, and attribution is required in customer applications.

    PricingUsage · from $4 1,000 requests · free tier
    TrustSOC 2 Type II · GDPR
    Does
    • AI answer
    • News search
    • Agent-ready / MCP
    • Zero data retention
    Used byApollo, AWS, Axel Springer, Chegg
    The catchAn independent index rather than a Google/Bing reseller, at a low per-query price, but results may not be stored, cached, or used to build a database.

    Brave Search API profile →

  • #4 Tavily Search API

    83 / 100

    Tavily Search API is a REST API built for AI applications that need real-time web access, offering search, content extraction, crawling, site mapping, and async deep-research tasks in a single service. It targets RAG pipelines, LLM grounding, and AI agents across 195+ countries, with SDKs for Python and JavaScript/TypeScript and an MCP server available. Pricing starts at $0.008 per credit with 1,000 free credits per month, self-serve signup, and an enterprise plan. The service is SOC 2 Type II certified and GDPR compliant, and its customer list includes AWS, IBM, Mastercard, MongoDB, and LangChain.

    PricingHybrid · from $0.008 credit · free tier
    TrustSOC 2 Type II · GDPR
    Does
    • Full content for RAG
    • Agent-ready / MCP
    Used byLangChain, MongoDB, Groq, Writer

    Tavily Search API profile →

  • #5 Linkup

    73 / 100

    Linkup is a production-grade web search API designed for AI applications, covering use cases from RAG and LLM grounding to deep research, legal AI, and sales enrichment. Pricing is usage-based starting at $0.001 per API call, with a $20 monthly free credit and self-serve signup, and an enterprise plan is available for higher rate limits and SLA coverage. The API ships SDKs for Python, JavaScript/TypeScript, and an MCP Server, with EU and US data residency options, SOC 2 Type 2 certification, and GDPR compliance. Notable customers include McKinsey, EY, Cohere, and GitGuardian.

    PricingUsage · from $0.001 API call · free tier
    TrustSOC 2 Type II · GDPR
    Does
    • Full content for RAG
    • Agent-ready / MCP
    • Zero data retention
    Used byMcKinsey & Company, Cohere, LEGORA, SNCF

    Linkup profile →

  • #6 Jina AI Search Foundation (DeepSearch / Reader)

    58 / 100

    Jina AI Search Foundation is a suite of LLM-oriented search and retrieval APIs covering URL-to-Markdown conversion, web search, iterative deep research, fact-checking, multilingual embeddings, reranking, and text segmentation, aimed at developers building RAG pipelines, AI agents, and answer engines. Pricing is usage-based at $0.05 per million tokens, with a 10 million token free tier shared across all APIs and no sales call required to get started. EU-dedicated endpoints are available for data-residency needs, the platform is GDPR-compliant, and a remote MCP server is supported alongside JavaScript and Node.js SDKs. API inputs and outputs are not used to train Jina's models.

    PricingUsage · from $0.05 1,000,000 tokens · free tier
    TrustGDPR
    Does
    • AI answer
    • Semantic / neural
    • Full content for RAG
    • Agent-ready / MCP
    Used bySingapore Airlines, Salesforce, Alibaba, AWS
    Avoid ifYou want to try it free before paying

    Jina AI Search Foundation (DeepSearch / Reader) profile →

  • #7 Parallel Search API

    61 / 100

    Parallel Search API is a REST-based web search and extraction service built for AI workloads, covering RAG grounding, agent pipelines, deep research, entity discovery, and web monitoring. Pricing is usage-based at $4.00 per 1,000 requests, with a free tier of 20,000 requests and an enterprise plan available; signup is self-serve. The API ships SDKs for Python, TypeScript/JavaScript, and LangChain, and offers an MCP server alongside webhooks. It holds SOC 2 Type 2 certification and contractually commits to not training on customer data, with zero data retention available under enterprise agreements.

    PricingUsage · from $4 1,000 requests · free tier
    TrustSOC 2 Type II
    Does
    • AI answer
    • Full content for RAG
    • Agent-ready / MCP
    • Zero data retention
    Used byClay, Sourcegraph, Harvey, Notion
    Avoid ifYou want to try it free before paying

    Parallel Search API profile →

  • #8 You.com Search API

    72 / 100

    You.com Search API provides real-time web search, news retrieval, and AI-grounded research capabilities designed for RAG pipelines, LLM agents, and agentic applications. Pricing starts at $5 per 1,000 searches on a usage-based model, with $100 in complimentary credits on signup and a free unauthenticated tier of 100 queries per day. The REST API supports Python and TypeScript SDKs, an MCP server, and both API key and OAuth2 authentication, and is SOC 2 Type 2 certified with GDPR compliance. Customers include Amazon, Databricks, DuckDuckGo, Bank of America, Harvey, and Windsurf.

    PricingUsage · from $5 1,000 searches · free tier
    TrustSOC 2 Type II · GDPR
    Does
    • AI answer
    • Full content for RAG
    • News search
    • Agent-ready / MCP
    Used byDuckDuckGo, Harvey, Windsurf, Databricks
    Avoid ifYou want to try it free before paying

    You.com Search API profile →

  • #9 Kagi Search API

    53 / 100

    Kagi Search API provides programmable access to Kagi's premium search index, covering web, image, news, video, podcast, and content extraction endpoints, with built-in summarization and citation-grounded answers suited for RAG pipelines and AI agents. Pricing is usage-based at $12 per 1,000 searches with no sales call required, and an enterprise plan is available. The API is GDPR compliant, supports Python, Go, Rust, and TypeScript SDKs, and ships an MCP server for agent tooling integrations.

    PricingUsage · from $12 1,000 searches · free tier
    TrustGDPR
    Does
    • AI answer
    • Full content for RAG
    • News search
    • Agent-ready / MCP
    Avoid ifYou want to try it free before paying

    Kagi Search API profile →

  • #10 Perplexity Sonar API

    73 / 100

    Perplexity Sonar API is a REST-based search and language model service that lets developers add real-time web retrieval, Q&A with inline citations, deep research, and semantic embeddings to their products. Pricing is usage-based starting at $1.00 per million tokens, with self-serve signup and no sales call required. The API ships SDKs for Python, TypeScript, and Node.js MCP, holds SOC 2 Type 2 certification, enforces a zero data-retention policy, and counts Samsung among its documented customers.

    PricingUsage · from $1 1M tokens · free tier
    TrustSOC 2 Type II
    Does
    • AI answer
    • Semantic / neural
    • Agent-ready / MCP
    • Zero data retention
    Used bySamsung
    The catchAn LLM answer engine priced per token, not a raw results API; ideal for cited answers, the wrong tool if you need ranked links to rank-track or scrape.

    Perplexity Sonar API profile →

  • #11 Mojeek Web Search API

    39 / 100

    Mojeek Web Search API provides programmatic access to an independent web index of over 9 billion pages, built for embedding search into applications, grounding LLMs and RAG pipelines, and powering AI agents. Pricing is usage-based per 1,000 queries with self-serve signup, tiered plans covering up to 400,000 queries per day, and custom enterprise arrangements. Mojeek is GDPR-compliant and has been used by Kagi and Meta's BlenderBot 3, giving it a track record in both commercial search products and large-scale AI research.

    PricingUsage · free tier
    TrustGDPR
    Does
    • Semantic / neural
    Used byKagi, SeeKeR (Meta/Facebook BlenderBot 3)
    Avoid ifYou want to try it free before paying

    Mojeek Web Search API profile →

  • #12 Google Programmable Search Engine (Custom Search JSON API)

    47 / 100

    Google Programmable Search Engine (Custom Search JSON API) lets developers embed Google web and image search results into applications via a REST API authenticated with an API key, with SDKs available for seven languages including Python, Java, and Go. Pricing starts at $5 per 1,000 queries, with a free tier of 100 queries per day and a hard ceiling of 10,000 paid queries per day. The API is closed to new customers, and existing users must migrate to an alternative solution by January 1, 2027.

    PricingUsage · from $5 1,000 queries · free tier
    The catchClosed to new customers (existing users must migrate by January 1, 2027); listed for reference, not a buildable option today.

    Google Programmable Search Engine (Custom Search JSON API) profile →