Lithuanian Pharma E-Shop platform with maximum of 10k items, including full text search, attribute filters, autocomplete and so on. Which of the following search engines would be the best fit for the job?

GPT-5.4

Vote: Meilisearch

For a catalog of only ~10k items, Meilisearch is the best fit because it provides excellent full-text search, typo tolerance, fast autocomplete, and faceted/attribute filtering with much less operational complexity than Elasticsearch or Solr. Elasticsearch and Solr are powerful, but they are usually better justified when scale, complex analytics, distributed indexing, or advanced relevance tuning are primary needs. For a small-to-medium e-commerce catalog, they can be overkill. MySQL full-text search is the weakest option here. It can work for basic keyword matching, but it is not the best tool for rich search UX such as autocomplete, typo handling, and flexible filtering. Meilisearch is especially attractive for a modern e-shop because it is easy to set up, fast out of the box, and designed for developer-friendly search experiences. For this size and feature set, it is the most practical choice.

Grok 4.1 Fast

Vote: Meilisearch

For a Lithuanian Pharma E-Shop with only 10k items, Meilisearch (D) is the best fit due to its lightweight design, instant setup, and perfect feature match: full-text search with typo tolerance, attribute filters/facets, and built-in autocomplete. ElasticSearch (A) and Solr (B) are enterprise-grade but overkill—high resource demands, complex clustering, and steep learning curves unnecessary for this scale. MySQL full-text (C) lacks robust facets, autocomplete, and advanced relevance scoring out-of-the-box, requiring custom hacks. Meilisearch delivers sub-50ms queries on modest hardware, easy Docker deployment, and e-commerce-ready APIs, making it ideal for quick MVP to production without bloat.

DeepSeek V3.2

Vote: Meilisearch

For a small-scale Lithuanian pharma e-shop with a maximum of 10k items, Meilisearch offers the best balance of features, ease of use, and low operational overhead. It provides built-in full-text search, attribute filtering, autocomplete, typo tolerance, and faceting out of the box, which are essential for an e-commerce search experience. While ElasticSearch and Solr are powerful, they are overkill for this dataset size and require more setup and maintenance. MySQL full-text search is limited in functionality and performance compared to dedicated search engines. Meilisearch's lightweight architecture and developer-friendly design make it the optimal choice for this scenario.

Gemini 3.1 Pro

Vote: Meilisearch

For a small dataset of only 10,000 items, heavyweight JVM-based search engines like Elasticsearch (A) or Solr (B) are significant overkill. They require substantial server resources, complex configuration, and continuous maintenance that aren't justified for this scale. On the other hand, MySQL full-text search (C) is too basic and lacks out-of-the-box features expected in modern e-commerce, such as robust typo tolerance, relevancy ranking, and instant autocomplete. Meilisearch (D) is the perfect fit here. It is a lightweight, Rust-based search engine specifically designed to deliver lightning-fast, highly relevant search experiences for datasets under 100k records. It offers out-of-the-box typo tolerance, seamless autocomplete, and easy-to-configure attribute filtering (faceting). This will save development time and infrastructure costs while providing a top-tier user experience for the e-shop.

Claude Opus 4.6

Vote: Meilisearch

For a Lithuanian Pharma E-Shop with a maximum of 10k items, Meilisearch is the best fit for several reasons: **Right-sized for the job:** 10k items is a small dataset. ElasticSearch and Solr are powerful but heavyweight solutions designed for millions or billions of documents. They come with significant operational complexity (JVM tuning, cluster management, etc.) that is overkill for this scale. **Built-in features out of the box:** Meilisearch provides full-text search, faceted/attribute filtering, typo tolerance, and autocomplete with virtually zero configuration. These are exactly the features required by the platform. With Elasticsearch or Solr, you'd need to configure analyzers, suggesters, and facets manually. **Typo tolerance and language handling:** Meilisearch has excellent built-in typo tolerance which is particularly valuable for a Lithuanian pharmacy platform where users may misspell drug names or Lithuanian words. It handles this gracefully without custom analyzer chains. **Ease of deployment and maintenance:** Meilisearch is a single binary with minimal resource requirements. For a small e-shop team, this dramatically reduces DevOps overhead compared to managing an Elasticsearch or Solr instance with its JVM dependencies. **Why not MySQL full-text search?** While the dataset is small enough for MySQL, MySQL's full-text search lacks proper autocomplete, typo tolerance, faceted filtering, and relevance tuning. It would require significant custom code to approximate the user experience that Meilisearch provides natively. **Performance:** Meilisearch is designed for instant search (<50ms responses) which is ideal for autocomplete in an e-shop. At 10k documents, it will feel instantaneous. Meilisearch is purpose-built for exactly this kind of use case: a small-to-medium catalog with search-as-you-type, filters, and a great user experience, all with minimal setup.

Final Summary

The AI Roundtable reached a rare moment of total alignment, with every model rejecting the industry giants in favor of a specialized, lightweight tool. Claude Opus 4.6 and Gemini 3.1 Pro led the charge, labeling traditional solutions like Elasticsearch and Solr as 'significant overkill' for a mere 10,000 items, while GPT-5.4 and Grok 4.1 Fast emphasized that operational simplicity is the ultimate feature for a small pharma e-shop.

No agreement after 1 rounds of discussion. Final vote: Meilisearch: 5

Strongest Arguments

• Meilisearch: Meilisearch is 'right-sized' for 10k items, offering native typo tolerance and instant search-as-you-type functionality without the massive hardware overhead or configuration complexity of enterprise-grade JVM engines.