Quickwit is still a young search engine but it already offers several features, here are the main ones:

• Full text search with phrase query and configurable tokenizers
• Natural query language: (michael AND jackson) OR "king of pop")
• Native support for time partitionning
• Local and remote (S3-like storage) indexes
• Distributed search with stateless instances
• Fast search on object storage or any storage which provides bytes range queries
• Fast startup time, indexing and search
• Mapping feature to define your schema and easily convert a json into a document to index
• Support data types text, i64, f64, date and composite types object and array.

You can have a look at tantivy features to dig into the indexing engine.

• no support for file formats other than ndjson
• no support for object storages not compatible with Amazon S3
• no faceted search
• no deletions (append mode only)

Influence our roadmap by voting in our GitHub issues for the features that you need the most!