Meilisearch is a fast, open-source search engine with a rich feature set including vector search and hybrid ranking. These exercises cover the configuration and query patterns you need to build production search experiences.
0 / 5 completed
1 / 5
At standup, a colleague asks what filterable attributes are in Meilisearch. What is the correct answer?
Filterable attributes are fields you configure in Meilisearch's index settings (filterableAttributes) so the engine builds a filter index for them at indexing time. At query time you use the filter parameter with expressions like brand = 'Nike' AND price < 100 to narrow results. Only fields declared as filterable can be used in filter expressions — using an undeclared field raises an error.
2 / 5
During a PR review, a teammate asks what sortable attributes configuration enables. Which answer is correct?
Meilisearch requires you to declare sortableAttributes in index settings before you can sort on those fields. At indexing time, Meilisearch builds a sort structure for declared fields. At query time you pass sort: ['price:asc', 'name:desc']. Fields not in sortableAttributes cannot be used in sort. Declaring a field sortable does not automatically make it filterable — those are separate settings.
3 / 5
In a design review, the team discusses vector search in Meilisearch. A junior engineer asks how it differs from keyword search. What is correct?
Meilisearch's vector search stores embedding vectors in a _vectors field and uses approximate nearest-neighbour (ANN) search to find semantically similar documents. You can use it standalone (pure semantic search) or combine it with keyword search as hybrid search, tuning the balance with the semanticRatio parameter (0 = pure keyword, 1 = pure vector). Embeddings can be generated by Meilisearch (via configured embedders like OpenAI) or provided by your application.
4 / 5
An incident report shows slow indexing after enabling multiple indexes with overlapping data. A senior engineer asks what multi-search does in Meilisearch. What is correct?
Meilisearch's multi-search endpoint (POST /multi-search) batches multiple search queries into a single HTTP request. Each query in the queries array specifies its own indexUid and search parameters, and results come back as an array in the same order. This is useful for federated search UIs (searching products, articles, and users in one request) and reduces network round-trips compared to making separate calls.
5 / 5
During a code review, a senior engineer asks what hybrid search in Meilisearch requires to be configured. What is accurate?
To use hybrid search in Meilisearch you must first configure an embedder in the index's embedders settings — specifying the source (openAi, huggingFace, rest, etc.), model, and credentials. Meilisearch then generates and stores vectors for new documents automatically. At query time you pass hybrid: { semanticRatio: 0.5, embedder: 'default' } to blend keyword and semantic scores.