Parseable
Rust-based log lake that writes Parquet to S3-compatible object storage. Compute-storage separated, indexless by design.
| License | AGPL-3.0 |
| Repository | https://github.com/parseablehq/parseable |
| GitHub stars (approx) | 2500 |
| Language | Rust |
| Self-host deployment | single binary, docker, kubernetes helm |
| Commercial hosted option | yes (vendor-run) |
| Freshness status | Fresh |
| Freshness checked | 2026-05-17 |
| Categories | logs |
| Replaces SaaS | datadog, splunk, new-relic |
Parseable is a logs-only backend built around the principle “object storage is the database.” Logs land as columnar Parquet on S3/MinIO/GCS; queries push down to the storage layer. No Elasticsearch-style cluster to babysit.
Pick it if: your log volume is large and bursty, your cost model is dominated by Elasticsearch JVM heap or ingestion-priced SaaS, and you already trust S3 (or MinIO) for durability. Schema-on-read is a feature for messy multi-tenant log streams.
Skip it if: you need full-text search at sub-second latency on hot data (OpenSearch is better) or you need a unified metrics/traces/logs pane — Parseable is logs-first and intentionally narrow. AGPL may also be a non-starter for some commercial embeds.
Deployment: single static binary or container; point at any S3-compatible bucket. Helm chart for k8s. The ingest API is JSON over HTTP — simple to integrate from Vector, Fluent Bit, or OTel.