Why I choose Meilisearch for frontend search
Meilisearch secured 15 million dollars in Series A funding to expand its enterprise-focused strategy and cloud offerings. The Rust-based search engine provides a developer-friendly alternative to Algolia and Elasticsearch for frontend applications requiring fast deployment.
Meilisearch grows its developer community
Meilisearch closed a $15 million Series A round led by Felicis to expand its marketing and sales teams. This capital helps the Paris-based company transition to an enterprise-focused strategy. CEO Quentin de Quelen, who co-founded the project in 2018 with Clement Renault and Thomas Payet, focuses on Meilisearch Cloud. With over 10 million downloads and more than 10,000 apps, the project has established a strong developer community. De Quelen created the search API based on his experience at Veepee and Louis Vuitton. He notes that building great search experiences usually requires large tech resources. Meilisearch avoids this difficulty by providing a solution for frontend applications across many domains. The tool uses natural language processing to understand user queries. The company plans to expand its 25-person headcount to 30 by the end of the year and to 50 by the end of 2023. This expansion follows the $15 million investment from Felicis and other partners like CRV and LocalGlobe. The total funding for Meilisearch has reached $22 million.
Comparing the search engine architecture
Developers weigh Meilisearch against Algolia and Elasticsearch based on data limits and deployment needs. Algolia works as a hosted search-as-a-service platform. It keeps indexes in RAM to deliver results in milliseconds. The software handles spelling mistakes and provides custom synonym mapping. Elasticsearch is an open-source engine built on Apache Lucene. It manages billions of records through shards and replicas. Because Meilisearch uses a self-contained binary with a built-in HTTP server, developers can deploy it quickly without the need to manage a complex backend like the JVM or a separate reverse proxy.
| Feature | Algolia | Elasticsearch | Meilisearch |
|---|---|---|---|
| Language | C++ | Java | Rust |
| Max Index Capacity | 128 GB | Unlimited | 500 GB |
| Max Record Size | 10 KB | Unlimited | 2 GB |
| Deployment | SaaS | Self-hosted or Cloud | Self-hosted or Cloud |
| High Availability | Yes | Yes | No |
Algolia caters to users who want fast implementation without managing infrastructure. It relies on a subscription model that charges based on records and operations. Because Elasticsearch gives developers full control, they must manage shards, replicas, and cluster health. Users can also use it for live logs and time-based data analysis. Meilisearch provides a similar managed experience to Algolia through Meilisearch Cloud. This cloud version onboarded 50 companies during its private beta. Meilisearch supports major languages and includes filters for price and date. It also corrects for typos to ensure errors do not impact the user experience. Algolia provides a dashboard that guides users through configuring index settings and uploading data. Elasticsearch is a tool you must deploy on servers or VMs in the cloud. You must configure the Java runtime, network ports, and security settings. It uses the BM25 algorithm by default, which ranks based on term frequency and uniqueness. Meilisearch provides a developer-friendly experience with a single binary. It manages ranking rules and supports custom filtering. Algolia is optimized for providing real-time results for typical user-facing searches. It is particularly favored by e-commerce and SaaS applications. Elasticsearch is effective for high-volume and intensive processing tasks, provided there are enough nodes. It can handle millions of documents every hour. Meilisearch’s performance is limited by available RAM. It suggests providing enough RAM to keep the entire dataset in memory for ideal performance. To upgrade Meilisearch, you must create a dump, replace the binary, and then load the dump. Meilisearch supports records up to 2 GB in size.
My verdict on Meilisearch
I recommend Meilisearch for developers who need a simple, powerful tool for frontend applications. It provides immediate utility without the steep learning curve found in Elasticsearch. You know how much time you waste configuring JVM settings just to get a basic search running. Meilisearch manages ranking rules and supports custom filtering. It also allows users to change ranking rules using a single index. However, Meilisearch only supports single-node settings. This means it lacks fault tolerance and creates a single point of failure in production. If you require high availability, you may need to look elsewhere. Will the company introduce multi-node support in its next version?