> The flexible AI-powered Search & Discovery platform
Offers "[[Serverless MOC|Serverless]]" PAYG pricing. Has 3 per-"unit" [pricing plans](https://www.algolia.com/pricing/):
- Up to 10 units
- Includes search with Analytics
- AI-powered search with advanced features including Personalization, Rules and Merchandising.
**1 unit** = 1,000 search requests AND THE CAPACITY TO INDEX UP TO 1,000 RECORDS
## Building with Algolia
- [[Integrating Algolia with your AWS serverless application]]
- How does the limit on indexing up to 1,000 records in the pricing description work?
- How could you put a caching proxy in front of Algolia (e.g. [[AWS CloudFront|CloudFront]]) in order to cache common searches? This could mitigate costs of these searches.
- When should you use a single index for multiple entities and when's it more beneficial to have each entity in their own? Similar to [[Single-table vs multi-table DynamoDB data models]] in DynamoDB - I expect the decision is to keep all entities in the same index which you may need to return in the same result set — which is dependent on the front-end use case.
- Can scoped tokens have a filter built into them?
tags: #Serverless #Search #ServerlessProduct