Modern products rely on search for e-commerce, documentation and AI assistants. API search pricing looks simple on the pricing page, but real cost shifts with traffic patterns, index size and AI features.
This guide explains how major providers charge, how site and app search differs from web SERP APIs and data APIs, and how to compare plans before you pick one.
How API Search Pricing Usually Works
Most vendors charge for a mix of three things: how many searches you run, how much data you index, and how much compute or infrastructure you use to serve and update that data.
On top of this, many now meter AI features such as semantic ranking, vector search and managed LLMs. Two plans that both say “from $30 per month” can end up very different once you hit millions of documents or start using heavy AI options. Always look at the base fee plus the extra cost per 1000 searches, per 1000 records or per GB stored, not just the headline price.
Site and App Search API Pricing
These APIs power product search, help centers and in-app search for logged in users. Some teams also use these APIs to turn company homepages and pricing pages into real-time competitive intelligence, not just to power on-site search.
Algolia
Algolia charges for both requests and records. The Grow plan includes about 10,000 searches and 100,000 records, then roughly $0.50 per extra 1000 searches and $0.40 per extra 1000 records. Grow Plus adds AI features and raises search cost to around $1.75 per 1000 searches, in exchange for better relevance tooling. Elevate is an annual enterprise plan with volume discounts, broader AI options and stronger SLAs.
Meilisearch
You can self host Meilisearch at no license cost and only pay for your own infrastructure, which suits teams that are comfortable running servers. Meilisearch Cloud offers managed usage based plans from about $30 per month, plus resource based plans priced by instance size and storage. Those resource plans start in the low tens of dollars per month, for example around $20 per month for 32 GiB base storage, with extra disk billed per GiB.
Elastic Cloud and Elasticsearch Serverless
Elastic Cloud Hosted runs managed clusters on tiers such as Standard, Gold and Platinum. Small Standard clusters are often quoted from around the mid $90s per month, with higher tiers adding machine learning features, security options and support.
Elasticsearch Serverless removes manual cluster sizing. You pay for ingest, search and machine learning in virtual compute units, storage per GB per month and any managed LLM usage per million tokens. This suits AI heavy or spiky workloads but requires monitoring so that surprise traffic does not create surprise bills.
Azure AI Search
Azure AI Search uses search units (SUs) across tiers from Free to Storage Optimized. In many regions, Basic is around $70–$80 per month and Standard S1 around $240–$250, with each tier setting storage limits, replica counts, and index caps. Features such as semantic ranker and agentic retrieval are billed separately after a small free allowance, on top of the base SU cost.
SeekStorm
SeekStorm provides search as a service with built in crawling. After a free trial, Basic is about $19 per month for one index, up to one million documents, 1 GB storage and around one million operations. Pro and Business raise limits to roughly 100 million documents and 100 million operations. Operations cover both indexing and queries, so you need a rough idea of how often your content changes.
Web and SERP Search API Pricing
These APIs return Google or Bing style results or scrape public sites for SEO tools, monitoring and AI agents, and are often provided by leading API search companies focused on SERP and web data.
SerpAPI
SerpAPI offers fixed monthly plans with defined searches and throughput. The free tier has 250 searches and paid plans start around $25 per month for 1000 searches, then scale up with lower per search prices as volume grows.
Bright Data
Bright Data prices Web Unlocker, Crawl API and SERP API mostly per 1000 successful results or requests. Pay as you go is about $1.5 per 1000, with large bundles such as two million results per month closer to $1 per 1000. Browser APIs and proxies are usually billed by GB or by IP.
Oxylabs
Oxylabs offers proxy products and a Web Scraper API starting around $49 per month. Residential proxies are typically billed per GB, often a few dollars per GB depending on volume and plan, and dedicated datacenter proxies are billed per IP, aimed at larger data collection stacks that run many parallel jobs.
Apify
Apify mixes plan credits with pay as you go compute, storage and proxies. A free plan includes a small credit. Paid plans around $29, $199 and $999 per month add more prepaid usage and lower unit prices, with compute units often near $0.3 each once you go beyond the included allowance.
Serper, Tavily, Exa and Zenserp
Newer APIs such as Serper, Tavily, Exa and Zenserp mostly sell credit bundles.
- Serper offers 50,000 credits for about $50, roughly $1 per 1000 queries, with cheaper rates at higher volumes.
- Tavily gives 1000 free credits per month, then about $0.008 per credit and small paid plans from roughly $30.
- Exa prices fast search at about $5 per 1000 operations with separate content and research pricing.
- Zenserp gives 50 free searches, then plans like $49.99 for 5,000 searches and $129.99 for 20,000, plus larger tiers.
Cost Comparison at a Glance
| Provider group | Typical starting point | What you mainly pay for |
| Algolia | Free tier, then Grow or Grow Plus | Requests, records and AI features |
| Meilisearch Cloud | Around $30 per month | Instance size and storage |
| Elastic Cloud Hosted | Mid $90s per month | Cluster resources and subscription |
| Elasticsearch Serverless | Usage only billing | Ingest, search, ML and storage usage |
| Azure AI Search | Around $73.73 per month per SU (Basic) | Search units by tier |
| SeekStorm | Around $19 per month | Documents and total operations |
| SerpAPI | Around $25 per month for 1000 searches | Number of SERP queries |
| Bright Data, Oxylabs, Apify | From tens of dollars per month | Results, requests, data or compute |
| Serper, Tavily, Exa, Zenserp | Credit bundles from about $50 | Credits or web search requests |
FAQs
What drives API search cost the most?
Search volume, indexed data and, at scale, AI and vector features.
Are free plans enough for production?
Usually not. They work for tests and very small apps. Most production products need a paid plan for capacity, reliability and support.
How do AI and vector features change pricing?
They add compute per query, so vendors charge higher per request rates or separate machine learning and token fees. That impact grows with traffic and index size.
Which pricing model is easiest to budget?
Fixed monthly or instance based pricing is simplest to predict. Usage based billing can be cheaper for variable traffic but needs monitoring and alerts.
What should I check before choosing a provider?
Look at how costs scale with traffic and data, overage rules, AI feature limits, SLAs, support, regions and how easy it is to switch plans later if your usage pattern changes.
