HelloStack
Exa logo

Search built for AI

Starting price
$0.012–$1.00per run
Pricing
Free plan
Website
exa.ai
Social media
@exaailabs

About Exa

Exa is a search engine built for AI that helps researchers find research papers on information retrieval. It serves developers and researchers who need a reliable API to access and extract web data for AI workflows.

Exa screenshot

Key Features

For researchers and developers, the value lies in a single API that unifies search, crawl, and semantic retrieval across multiple data sources, making AI workflows simpler:

Search

Query a list of results and their contents via a single API, returning results with embedded data.

Crawl

Retrieve all pages on a site and obtain their contents.

Answer

Generate natural-language summaries of web search results.

In-house vector database

Stores embeddings and enables fast semantic retrieval across data sources.

Summary

Best for AI researchers, ML engineers, and product teams building AI-powered search and retrieval workflows.

Plans and prices as listed on exa.ai as of Sep 24, 2026.

Starter

$0 / monthly

  • Free $10 credits per month
  • MCP server access
  • Claude Connector
  • ChatGPT plugin
  • 50+ integrations
  • Access to all endpoints
  • 10 Search Queries Per Second (QPS)
  • 50 Agent concurrency

Developer

$0 / pay as you go

  • Everything in starter
  • PAYGO usage-based billing
  • Standard email support
  • Pay via card or bank account
  • SOC 2 Type II
  • Up to 25 Search QPS
  • 50 Agent concurrency

Enterprise

$0 / custom

  • Everything in developer
  • Custom MSA and DPA
  • Zero data retention
  • SOC 2 Type II and HIPAA
  • Monthly invoicing
  • SSO and SCIM for Exa Dashboard
  • Organization unified billing
  • Enterprise support with dedicated Slack channel
  • Enterprise SLAs
  • Custom Search QPS & Agent Concurrencies
  • Up to 1,000 results per search

Agent

$0.012–$1.00 / per run

  • Asynchronous deep research runs
  • Structured outputs with citations
  • Access to leading data providers

Contents

$1 / per 1k pages

  • Full page contents for LLM context
  • Token efficient highlights
  • Configurable livecrawl policies

Search

$7 / per 1k requests

  • Web search tool calls for agents
  • Webpage text and highlights
  • Configurable latency: 180ms to 1s

Deep Search

$12–15 / per 1k requests

  • Multi-step agent workflows
  • Answers with structured outputs
  • Web grounded citations

Monitors

$15 / per 1k requests

  • Identify fresh events on the web
  • Runs searches at a specified cadence
  • Receive updates with webhooks

More in RAG

See all →