APIPriceBook

Search APIs: pricing compared

3 providers · normalized per 1,000 searches / capacity · June 2026 snapshot

Search APIs power site and in-app search. They split into two pricing models: usage-based (per search request and per record, like Algolia) and capacity-based (per provisioned node, like Typesense). The cheapest model depends on whether your query volume or your data size is the bigger driver. On its headline metric, Typesense Cloud is generally the cheapest (From ~$0.05 per node-hour (capacity-based, not per-search)), but the best value depends on your volume and feature needs.

Source: Vendor published pricing pages. Data as of June 2026.

Search APIs compared

Normalized per 1,000 searches / capacity.
ProviderHeadline priceFree tierBest for
Algolia~$0.50 per 1,000 search requests + record costsBuild plan: free up to 10,000 search requests and 1M records/month.Teams that want hosted, low-latency search-as-a-service with great relevance tooling out of the box.
Typesense CloudFrom ~$0.05 per node-hour (capacity-based, not per-search)Open-source self-host is free; Cloud is paid by node hour.Teams that want Algolia-like search at predictable capacity pricing, or to self-host for free.
Meilisearch CloudFree tier, then usage-based by documents + searchFree Cloud tier for small projects; open-source self-host is free.Developers who want a dead-simple, open-source search engine with an easy hosted option.

Vendor pricing as of June 2026 — verify on the vendor pricing page. Usage-based vendors charge per 1,000 search requests plus a per-record fee. Capacity-based vendors charge per node-hour regardless of query count, which can be far cheaper at high query volumes but wasteful at low ones. Open-source engines (Typesense, Meilisearch) are free to self-host.

How we compare

Usage-based vendors charge per 1,000 search requests plus a per-record fee. Capacity-based vendors charge per node-hour regardless of query count, which can be far cheaper at high query volumes but wasteful at low ones. Open-source engines (Typesense, Meilisearch) are free to self-host.

Provider notes

Algolia — ~$0.50 per 1,000 search requests + record costs

The hosted search standard. Priced per search request AND per record, so both query volume and catalog size drive cost. Excellent DX and relevance controls.

Free tier: Build plan: free up to 10,000 search requests and 1M records/month.

Typesense Cloud — From ~$0.05 per node-hour (capacity-based, not per-search)

Open-source alternative to Algolia. Because you pay for capacity (node-hours) rather than per search, high-query apps can be far cheaper, but you size and manage the cluster.

Free tier: Open-source self-host is free; Cloud is paid by node hour.

Meilisearch Cloud — Free tier, then usage-based by documents + search

The simplest of the open-source search engines to start with. Free to self-host; Cloud pricing scales with usage. Lighter on advanced relevance/merchandising than Algolia.

Free tier: Free Cloud tier for small projects; open-source self-host is free.

Head-to-head matchups

Frequently asked questions

What is the cheapest search API in 2026?

Typesense Cloud is generally the cheapest search option on its headline metric (From ~$0.05 per node-hour (capacity-based, not per-search)). The best value depends on your volume and feature needs — Usage-based vendors charge per 1,000 search requests plus a per-record fee. Capacity-based vendors charge per node-hour regardless of query count, which can be far cheaper at high query volumes but wasteful at low ones. Open-source engines (Typesense, Meilisearch) are free to self-host.

How do you compare search API pricing fairly?

We normalize each provider to a common unit (per 1,000 searches / capacity). Usage-based vendors charge per 1,000 search requests plus a per-record fee. Capacity-based vendors charge per node-hour regardless of query count, which can be far cheaper at high query volumes but wasteful at low ones. Open-source engines (Typesense, Meilisearch) are free to self-host. Always confirm current pricing on the vendor's page, since rates change.

Source & accuracy

Each figure is a snapshot of the named vendor's public pricing page, captured in June 2026. Trademarks belong to their owners. API pricing changes frequently — verify on the vendor's pricing page before relying on these numbers. Derived figures are computed with the transparent formulas described in our methodology.

Last updated: 2026-06-18