Exa
Exa is an AI tool designed to connect other AI systems to the internet, with a primary focus on improving search capabilities of AI. Exa AP...
Last verified:
What is Exa?
Exa is a search engine built specifically for AI applications, designed from scratch to find high-quality information for LLMs. It powers thousands of companies' LLM and agentic applications by providing real-time web search with structured outputs, full page content extraction, and token-efficient highlights specifically optimized for AI workflows.
Key features include semantic embedding search using transformers, natural language query processing, instant content retrieval in clean markdown format, sub-200ms low-latency search options for agent loops, automatic search method selection (auto type), deep search with reasoning capabilities, structured JSON outputs via schema, content highlights that reduce token budgets by 50%, subpage crawling for entire site sections, and specialized web indexes for people, companies, code docs, financial data, and news. The Contents API extracts full page text, targeted highlights, and LLM-generated summaries from any URL while handling JavaScript-rendered pages and PDFs automatically.
Exa is designed for developers building AI applications, coding agents that need to search GitHub repos and documentation, research automation tools, LLM-powered chatbots requiring grounded web citations, startups and education projects through free grants, and enterprise teams needing zero data retention and custom indexes. It's used by companies like Cursor for accurate code documentation, OpenRouter for model-agnostic grounding, and StackAI for powering AI agent workflows.
Exa pricing
Pricing model: Freemium
Free tier: 1,000 requests per month for free plus $10 initial credits for new users. Startup and Education Grants: $1,000 worth of free credits. Search endpoint: $7 per 1k requests (up to 10 results), $1 per additional result. Deep Search: $12 per 1k requests. Deep-Reasoning Search: $15 per 1k requests. Contents: $1 per 1k pages per content type (text, highlights, or summaries). Monitors: $15 per 1k requests for tracking web events. Answer endpoint: $5 per 1k requests for fast web-grounded answers with streaming. AI page summaries: $1 per 1k pages across all endpoints. Enterprise: Custom pricing with volume discounts, custom rate limits, SLAs, and 1:1 onboarding. Websets Free: 1,000 credits with up to 25 results. Websets Core: $49/month for 8,000 credits. Websets Pro: $449/month for 100,000 credits.
Exa pros
- Built from scratch specifically for LLM applications, not adapted from traditional search
- No ad revenue incentive means fully focused on returning highest quality results
- Embedding-based semantic search understands meaning, not just keywords
- Sub-200ms latency options ideal for latency-sensitive agent loops
- Token-efficient highlights reduce LLM costs by over 50%
- Structured JSON outputs via schema for complex enrichment workflows
- Search over 70M+ companies with extracted structured data like CEO and founded year
- Dedicated web indexes for people, companies, code docs, financial reports, and news
- Automatically handles JavaScript-rendered pages, PDFs, and complex layouts
- Subpage crawling pulls content from entire site sections in single request
- Index updated every hour with constantly adding new links
- Up to 1,000 results per search with options for thousands in batch use cases
- $1000 free credits available for startup and education grants
- Zero data retention available as enterprise security feature
- Natural language queries instead of requiring complex query syntax
- Similarity search finds pages talked about in similar ways to URL input
- Streaming responses with web grounded citations for Answer endpoint
- MCP setup available to connect AI assistants directly to Exa
Exa cons
- No permanently free tier, only initial trial credits for evaluation
- $10 initial credits deplete quickly for production workloads
- Pricing jumps to $49/month for 8,000 credits on Websets Core plan
- Additional results above 10 cost $1 per 1k requests on top of base price
- Deep-Reasoning Search at $15 per 1k requests is expensive for complex queries
- Free tier intended for evaluation only, not fixed monthly allowance
- Custom index and zero data retention require enterprise agreement
- Limited coverage for domains not in client's specific use case without custom arrangement
- 10 credit = 1 all-green result in generated webset can add up quickly
- Does not crawl pages with noindex tag like any search engine
Frequently asked questions about Exa
What is Exa?
Exa is a search engine built specifically for AI applications. It was built from scratch and is state of the art at finding high quality information for LLMs. Exa is used by thousands of companies to power their LLM and agentic applications, enabling developers to filter the internet's content with complex queries to bring world knowledge into their AI apps.
What's different about Exa Search compared to traditional search?
Traditional search engines are optimized for clicks and ads, and nearly every search API wraps traditional search engines with similar problems. In contrast, Exa is optimized to return the highest quality information for LLM applications. Exa does not make money from ads, so it is fully incentivized to return highest quality results. Because Exa built its own search engine from scratch, it provides customized features other providers cannot offer.
How is Exa different from LLMs?
Exa is a search engine built from the ground up, while LLMs are models built to predict the next piece of text. Exa predicts specific links on the web given their relevance to a query. LLMs have intelligence and get smarter as new models are trained, while Exa connects these intelligences to the web by providing access to current information.
How can Exa be used in an LLM?
Exa enhances LLMs by supplying high-quality, relevant web content, minimizing hallucination and outdated responses. An LLM can take a user's query, use Exa to find pertinent web content, and generate answers based on reliable, up-to-date information from the web rather than relying solely on training data.
How does Exa compare to other search APIs?
Exa offers unique capabilities including embedding search technology using transformers for semantic understanding, natural language queries for accurate results, instant content retrieval with clean parsed content, large-scale searches capable of returning thousands of results, content highlights for targeted information, optimization specifically for AI applications and chatbots, auto search that automatically selects the best method, and constant index updates every hour.
How often is Exa's index updated?
Exa updates its index every hour and is constantly adding batches of new links. The system targets the highest quality web pages. Clients can request specific domains to be more deeply covered - if there's a use case that can be unlocked by additional domain coverage in the index, customers should contact Exa.
How does similarity search work in Exa?
When searching using a URL, Exa crawls the URL, parses the main content from HTML, and searches the index with that parsed content. The model chooses webpages it predicts are talked about in similar ways to the prompt URL, considering factors like text style, domain, and main ideas inside the text. This is a natural extension for a neural search engine and difficult with traditional search engines.
What security measures does Exa take?
Exa has robust policies with everything in standard cloud services or built in-house, including their own vector database served in-house, own GPU cluster, own query model, and own search solution. Exa can offer unique security arrangements like zero data retention as part of a custom enterprise agreement for enhanced privacy and compliance.
Does Exa have a crawler?
Yes, Exa crawls pages on the web just like any other search engine. However, if a webpage has the noindex tag and is therefore not crawlable by any search engine, then Exa will not crawl that page, following standard search engine practices.
What is on Exa's roadmap?
Exa's roadmap includes building a much larger index and solving search more comprehensively. The company continues to expand capabilities for AI applications, with ongoing development of features like Websets for sourcing knowledge workers, enhanced structured outputs, and improved coverage for specific use cases like people, companies, and code documentation.