Algolia pricing explained: what search-as-you-type does to usage
Understand Algolia search-request and record pricing, calculate a realistic catalogue example, and decide when hosted or owned search is the better fit.
Algolia pricing has two moving parts: search requests and records. The model is easy to underestimate because a search-as-you-type interface can send a request on every keystroke, while replicas and indexed objects affect record volume. The right estimate starts with product behaviour, not monthly active users.
What Algolia currently includes
Algolia's published Grow and Grow Plus plans include 10,000 search requests and 100,000 records each month. Grow lists additional search requests at $0.50 per 1,000 and Grow Plus at $1.75 per 1,000. Additional records are listed at $0.40 per 1,000 on both plans.
A worked catalogue example
Take a production catalogue with 500,000 records and 2 million search requests in a month. After the included allowance, Grow's published usage rates produce about $1,155 for that month: $995 for excess requests and $160 for excess records. Grow Plus produces about $3,642.50 at its published request rate. Enterprise discounts, contracts, replicas, and other products can change the actual bill.
Why keystrokes matter
Algolia's own pricing FAQ says autocomplete and search-as-you-type can trigger a request on every keystroke. Measure requests in production or replay representative sessions before committing to a forecast. Counting one search box submission per visitor will understate usage when the interface queries continuously.
Hosted Algolia or an owned index?
- Choose Algolia when fast implementation, its relevance tooling, and its managed global infrastructure matter more than variable usage cost
- Choose an owned search layer when the index is sensitive, query volume is predictable but high, or search behaviour is tightly coupled to your product
- Include operations, relevance tuning, failover, and analytics in the owned-search cost; servers alone are not a fair comparison
Atlas is ResoluteX's owned-search path for teams that want the index and query data inside their infrastructure with a flat commercial model. The useful comparison is not free versus paid. It is managed convenience versus control, including the labour required for each.
Continue the decision
Have the problem this post describes? The AI Architect will match it to a product or practice — with a price.
Run the AI Architect