Quickwit

Quickwit is an open-source search engine built for logs and traces. Its defining property is where the search happens. Indexes live in object storage, and stateless searchers query them in place, so there is no cluster of index nodes holding the only copy of the data.

Why it qualifies

• Object storage is the source of truth. Indexes are stored on object storage (S3, GCS, Azure, or MinIO), not on disks attached to specific nodes.
• Storage and compute are separate. Searchers hold no durable state, so they can be scaled up, scaled down, or replaced without moving data.
• Operational workload. It serves live search over logs and traces for observability.

Where it sits in the category

Quickwit is the open-source counterpart to turbopuffer, aimed at log and trace search. It stores its index in its own split format rather than an open table format like Parquet, so it meets the object-storage and separation criteria cleanly and is looser on the open-format one. Because it is open source, the format and the engine are both there to read.