Blog

August 12, 2026

Exa Pricing Guide: Understanding Costs for AI Search

TLDR: Exa uses a pure pay-as-you-go pricing model with no subscription and no minimum spend. You load credits and are billed per API request at rates that vary by endpoint and search depth. As of September 2026, standard search costs $7 per 1,000 requests, content retrieval costs $1 per 1,000 pages per content type, and deep research search starts at $12 per 1,000 requests. This article covers how every endpoint is priced, how to estimate costs for realistic workloads, and how Exa compares to Tavily and You.com.

How Exa's Pricing Model Works

Exa's pricing is straightforward in structure: there is no monthly plan, no seat license, and no tiered subscription. You add credits to your account and are charged per request at the rates listed below. All prices shown are from exa.ai/pricing, verified by live crawl on 2026-09-04.

Three concepts govern how your balance is consumed:

Request: one API call to an endpoint. Prices are quoted per 1,000 requests, so a $7 / 1k rate equals $0.007 per call.
Result: one search result returned in a response. The first 10 results in a request are covered by the base price. Results beyond 10 are billed at $1 per 1,000 additional results.
Page and content type: a page is one URL that Exa returns content for. A content type is one view of that page (text, highlights, or summary). The /contents endpoint bills each content type separately, so fetching text and highlights for the same URL counts as two pages billed.

This granularity means your actual cost depends on exactly which features you use per call, not just how many calls you make.

Pricing by Endpoint (as of September 2026)

Search API: $7 / 1k Requests

The core /search endpoint costs $7 per 1,000 requests. That base price covers up to 10 results per request. If your application requests more than 10 results, you pay an additional $1 per 1,000 excess results.

AI-generated page summaries, when requested, cost an extra $1 per 1,000 pages. This is billed regardless of which endpoint triggers the summary.

Deep Search: $12 to $15 / 1k Requests

Deep Search is activated by setting the type field on /search. It replaces the standard search with a multi-step research mode that produces structured outputs and citations. You.com offers a comparable capability through its deep research API if you are benchmarking multi-step synthesis across providers. There are three depth levels:

Type	Price	Typical latency	Best for
`deep-lite`	$12 / 1k requests	~4 seconds	Lightweight synthesis
`deep`	$12 / 1k requests	4 to 15 seconds	Multi-step reasoning with structured outputs
`deep-reasoning`	$15 / 1k requests	12 to 40 seconds	Harder research tasks

Additional results above 10 and AI page summaries are billed at the same rates as standard search when using Deep Search.

Contents API: $1 / 1k Pages Per Content Type

The /contents endpoint retrieves full page text, highlights, or AI-generated summaries for known URLs you already have. Unlike search, this endpoint does not find URLs for you. It is billed at $1 per 1,000 pages, per content type requested.

That per-type billing is worth understanding explicitly. If you request both text and highlights for the same URL in one call, that counts as two pages billed, not one. AI summaries add another $1 per 1,000 pages on top. A single URL requested with all three content types costs $0.003 in contents charges alone, before any search cost.

Answer API: $5 / 1k Requests

The /answer endpoint takes a question and returns an LLM-generated answer with citations. At $5 per 1,000 requests ($0.005 per call), it is cheaper than standard search on a per-call basis. It does not support additional results or AI summaries, so the price is flat. This endpoint is appropriate when your application needs a direct answer rather than a list of sources to process.

Monitors API: $15 / 1k Requests

Monitors are scheduled searches that surface new web events matching a query. They cost $15 per 1,000 scheduled runs, with the same $1 / 1k additional results and $1 / 1k AI page summaries as standard search. Per request, that ties with deep-reasoning search as the most expensive non-Agent endpoint.

Agent API: $0.012 to $1.00 Per Request (Fixed), or Usage-Based

The Agent API performs async deep research, list building, and enrichment. It can be called with a fixed effort level for predictable per-request pricing, or in a metered auto mode that bills by actual resource consumption:

Fixed effort level	Price per request
`minimal`	$0.012
`low`	$0.025
`medium`	$0.10
`high`	$0.50
`xhigh`	$1.00

Metered modes bill actual usage up to a cap. The auto mode defaults to a $5 cap per run. The beta max mode defaults to a $20 cap per run. Usage components in metered mode are:

Agent Compute Units (ACUs): $0.10 per ACU
Search tool calls: $0.005 per search
Email contact enrichment: $0.02 per email
Phone contact enrichment: $0.07 per phone number

Free Tier and Account Credits

New Exa accounts receive $20 in free credits, which Exa estimates covers approximately 2,800 standard searches at the $7 / 1k rate. After that, a Free Tier adds $10 in credits to your account every month automatically. There is no subscription required to receive the monthly $10 refresh.

This means a developer using only the free credits gets roughly $10 of search capacity each month indefinitely, which covers about 1,400 standard searches at the base rate.

Exa states explicitly: no subscription, no minimum spend. You can add credits as needed and pay nothing if you have remaining free balance.

Enterprise Pricing

Exa offers an enterprise tier for high-volume use cases, custom data requirements, and stricter security needs. As listed on the Exa pricing page (2026-09-04), the enterprise tier includes:

Up to 1,000 results per search request (the base price on other tiers covers 10 before extra-results billing)
Custom search QPS and Agent concurrency limits
Custom, tailored indexes and bring-your-own data providers
Enterprise SLAs, plus a custom MSA and DPA
Dedicated Slack channel support
Zero data retention
Monthly invoicing, SSO and SCIM, and unified organization billing
SOC 2 Type II and HIPAA, per Exa's own listing (verify current attestations with Exa directly)

Enterprise pricing is custom. Contact Exa directly for a quote.

Worked Cost Examples

These examples use verified September 2026 prices. They assume standard search with up to 10 results per request and no AI summaries, unless otherwise stated.

Example 1: Small Agent Doing 100 Searches Per Day

A lightweight research agent runs 100 basic searches per day:

Daily: 100 searches x $0.007 = $0.70
Monthly (30 days): $21.00

If you also fetch page text from 3 URLs per search using the Contents API:

Contents cost: 100 x 3 pages x 1 content type x $0.001 = $0.30/day, $9.00/month
Total monthly: $30.00

The $10 monthly free credit covers about half the raw search cost. At this scale, Exa's PAYG model costs less than many subscription tiers from other providers.

Example 2: Production App at 10,000 Searches Per Day

A production app running 10,000 standard searches per day with 5 results each (all within the 10-result base):

Daily: 10,000 x $0.007 = $70
Monthly: $2,100

Adding Contents API for 2 URLs per search at text-only:

Contents: 10,000 x 2 x $0.001 = $20/day, $600/month
Total monthly: $2,700

At this volume, it is worth contacting Exa about enterprise volume discounts.

Example 3: Deep Research at 500 Requests Per Day

A synthesis tool using deep search type at $12 / 1k requests:

Daily: 500 x $0.012 = $6
Monthly: $180

If 50 of those per day use deep-reasoning at $15 / 1k instead:

450 deep at $0.012 = $5.40/day
50 deep-reasoning at $0.015 = $0.75/day
Daily total: $6.15, monthly: $184.50

Example 4: Agent API for Lead Enrichment (Fixed Effort)

An agent enriching 200 leads per day using the medium fixed effort ($0.10/request):

Daily: 200 x $0.10 = $20
Monthly: $600

Switching to low effort ($0.025) where the task permits:

Daily: 200 x $0.025 = $5
Monthly: $150

Effort level selection has a 4x impact on Agent API cost in this example.

Cost Optimization Strategies

A few patterns reduce Exa costs materially without changing what your application does:

Request only what you use from Contents. Since each content type (text, highlights, summary) is billed independently, do not request all three if your pipeline only consumes one. Switching from requesting all three types to text-only cuts contents cost by two-thirds.
Limit result count when you do not need 10 results. If your agent typically uses 3 results, requesting the default 10 does not cost extra. But if your app is already requesting more than 10 and only using the first few, reducing numResults keeps you in the base-price tier.
Use the Answer API for simple factual queries. At $5 / 1k vs. $7 / 1k for Search, the Answer endpoint is cheaper per call for questions that do not need a result list. For complex queries where you need sources to inspect, Search is more appropriate.
Prefer fixed effort on Agent API over metered auto for predictable tasks. If your agent task is well-defined and consistently matches a known effort level, the fixed-effort tiers are cheaper than the metered default when your actual consumption is low in a metered run. Metered runs only save money when actual consumption is well below the cap.
Cache search results for repeated or near-identical queries. This applies to all search APIs. A simple local cache or a short TTL Redis layer can dramatically reduce calls for common query patterns in a production agent.

How Exa Pricing Compares to Tavily and You.com

For developers choosing between these providers, pricing model matters as much as per-unit price. Here is a direct comparison using verified September 2026 data:

Factor	Exa	Tavily	You.com
Pricing model	Pure PAYG, no subscription	Subscription tiers or PAYG at $0.008/credit	PAYG on prepaid credits, no subscription. $5/1k web search calls
Minimum monthly spend	None	None (free tier available)	None
Free allocation	$20 signup + $10/month recurring	1,000 credits/month, no card required	100 queries/day via the free MCP profile (no signup), plus $100 in credits on account signup
Base search cost	$0.007 per search	$0.008/credit x 1 credit = $0.008 (PAYG), lower on plans	$0.005 per search, up to 100 results. Full-page extraction adds $1/1k pages crawled live (cached pages free)
Deep research / synthesis	$0.012 to $0.015 per request (Deep Search)	$0.032 to $2.00 per request (Research, 4 to 250 credits depending on model and task, at $0.008 PAYG), lower on plans	Research API: $12 (lite) to $1,200 (frontier) per 1k calls (you.com/pricing, 2026-09-04)
Enterprise compliance	SOC 2 Type II, HIPAA, and zero data retention listed on the enterprise tier	Enterprise plan (custom)	Zero Data Retention available for the Web Search and Answer APIs (docs)
Predictable monthly cost	No cap, entirely usage-based	Yes: subscription plan sets a credit ceiling	Prepaid credits set a hard ceiling. Usage stops when the balance runs out unless you top up

Tavily's PAYG rate of $0.008 per credit is slightly higher than Exa's $0.007 per search at equivalent basic depth. Tavily's subscription plans reduce the per-credit cost to between $0.005 and $0.0075 depending on tier, making them competitive at committed volumes. Exa's enterprise volume discounts serve a similar purpose. For a side-by-side breakdown of how these billing models compare in practice, see the Tavily vs Exa comparison.

The core tradeoff: Tavily's subscription model gives you a predictable monthly ceiling and may be cheaper at high committed volume. Exa's PAYG model carries no risk of wasted spend from unused credits and no monthly minimum, making it more efficient for variable or unpredictable workloads.

Tavily Pricing Reference (as of 2026-09-04)

For direct comparison, Tavily's plans as published on docs.tavily.com/documentation/api-credits:

Plan	Credits / month	Monthly price	Effective cost per basic search
Researcher (free)	1,000	Free	Free (within limit)
Project	4,000	$30	$0.0075
Bootstrap	15,000	$100	$0.0067
Startup	38,000	$220	$0.0058
Growth	100,000	$500	$0.005
Pay As You Go	Per usage	$0.008 / credit	$0.008

Tavily advanced search (2 credits) effectively doubles the per-search cost on any plan. The Research endpoint has a dynamic model with per-request costs of 4 to 250 credits depending on model and task complexity: mini runs 4 to 110 credits per request and pro runs 15 to 250, per the same Tavily page.

You.com as an Alternative

You.com offers a web search API, Contents API, Research API, and Finance Research API. New accounts on the paid platform start with $100 in complimentary credits, more than Exa's $20 signup grant. A free MCP endpoint at https://api.you.com/mcp?profile=free provides access to you-search at 100 queries per day with no API key or signup, which is a useful starting point for developers evaluating the platform.

You.com Web Search costs $5 per 1,000 calls, with up to 100 results per call, and full-page extraction is a $1 per 1,000 pages add-on billed only on pages crawled live, so pages served from cache add nothing. The Contents API is also $1 per 1,000 pages. The Research API uses a tiered model keyed to the research_effort parameter: lite at $12, standard at $50, deep at $100, exhaustive at $450, and frontier at $1,200 per 1,000 calls. Current pricing is at you.com/pricing and API documentation is at you.com/docs (verified 2026-09-04).

Verifying Current Pricing

All prices in this article were verified by live crawl of exa.ai/pricing on 2026-09-04. API pricing changes regularly. Before building a production cost model, re-read the Exa pricing page directly. Exa also publishes a llms.txt index at https://exa.ai/docs/llms.txt and https://docs.exa.ai/llms.txt for programmatic documentation access if you want to track pricing changes.

If your monthly spend will exceed a few hundred dollars, contact Exa about enterprise terms. Volume discounts are available and are not listed publicly.