FluxOS aims to replace Postgres + Kafka + Flink + Materialize with a single Rust process. Everything below runs in flux-server, with no add-ons or agents. FluxOS is in private beta: the parts marked experimental work but have known gaps, listed on the known limits page.
ACID, MVCC snapshot isolation, real BEGIN / COMMIT / SAVEPOINT. Self-hosted servers speak the Postgres wire protocol on port 5432 for psql and drivers, with known gaps; Cloud workspaces don't yet, and most ORMs don't work yet (known limits).
Continuous queries are materialized views that stay fresh as data arrives, with windowing, ASOF joins and watermarks in-process. Kafka and Kinesis sources and sinks are experimental.
Columnar SSTs with ZSTD compression, time-travel temporal queries, SCD-2 history, 155+ built-in functions, window functions, semantic auto-MVs. The dashboard reads from the same tables the app writes to. No ETL. No reverse ETL.
Each item below maps to a source file and tests in the codebase. Items marked experimental have known gaps.
Self-hosted servers accept psql and Postgres drivers on port 5432 (with known gaps). Not on Cloud yet, and most ORMs don't work yet.
Transactions with snapshot isolation, not eventual consistency.
Kafka + Kinesis sources and sinks. Windows. ASOF joins. Watermarks. In-process.
Streaming MVs that update as data arrives. The Materialize feature, without the separate cluster.
An event stream of every row change. Webhooks. Cursor checkpointing. Replay. A destination that falls far behind can miss changes.
Query a past point in time, with built-in slowly-changing-dimension history. After compaction, older points can return wrong results; see known limits.
Leader election with epoch fencing and WAL replication. Self-hosted only; not offered on Cloud.
Hot WAL + catalog stay local. Cold SSTs land in S3 (or any S3-compatible store).
Roles, permissions, token TTL, brute-force lockout, password policy. No external IdP required.
An append-only NDJSON record of who did what. Its hash chain can't be verified yet.
gdpr erase finds and removes a person's rows across tables by column value. Self-hosted only for now.
AES-256-GCM on self-hosted servers. Cloud workspaces use storage-level encryption for now.
You can try these today, but they have known correctness gaps. Don't rely on them for data you need.
Creating a branch from a branch, and merging, have known data-loss bugs.
Iceberg and Delta tables.
Kafka and Kinesis sources and sinks.
train model, predict, and vector indexes.
Natural-language insights from an LLM you connect.
Installable packages, and mounting another FluxOS server's tables.
Replication and failover between self-hosted nodes.
Per-warehouse resource limits.
FluxOS is invite-only while we harden the core, and free during the beta.