OpenData is a family of open-source, object-native databases that reduces costs, simplifies operations and scales with ease.

Databases made simple

All OpenData systems are built on the same foundation so once you know how to operate the first, you know how to operate the rest.

  ╔══════════════╗  ╔══════════════╗  ╔══════════════╗  ╔══════════════╗
  ║              ║▌ ║              ║▌ ║              ║▌ ║              ║▌
  ║    Vector    ║▌ ║    Buffer    ║▌ ║  Timeseries  ║▌ ║     Log      ║▌
  ║              ║▌ ║              ║▌ ║              ║▌ ║              ║▌
  ╚═══════┬══════╝▌ ╚═══════┬══════╝▌ ╚═══════┬══════╝▌ ╚═══════┬══════╝▌
   ▀▀▀▀▀▀▀│▀▀▀▀▀▀▀▘  ▀▀▀▀▀▀▀│▀▀▀▀��▀▀▘  ▀▀▀▀▀▀▀│▀▀▀▀▀▀▀▘  ▀▀▀▀▀▀▀│▀▀▀▀▀▀▀▘
          │                 │                 │                 │
          └─────────────────┴────────┬────────┴─────────────────┘
                                     │
                                     │
                                     │
            ┏━━━━━━━━━━━━━━━━━━━━━━━━▼━━━━━━━━━━━━━━━━━━━━━━━┓
            ┃                                                ┃
            ┃            OpenData Storage Engine             ┃
            ┃                                                ┃
            ┃                built on SlateDB                ┃
            ┃                                                ┃
            ┗━━━━━━━━━━━━━━━━━━━━━━━━┬━━━━━━━━━━━━━━━━━━━━━━━┛
                                     │
                                     │
                                     │
          ╔══════════════════════════▼═════════════════════════╗
          ║                                                    ║
          ║         Object Storage   ·   S3 · GCS · R2         ║
          ║                                                    ║
          ╚════════════════════════════════════════════════════╝

╔══════════════╗  ╔══════════════╗  ╷   ┏━━━━━━━━━━━━━━━━━━━┓
║              ║  ║              ║  │   ┃     OpenData      ┃
║    Vector    ║  ║    Buffer    ║  ├──►┃  Storage Engine   ┃
║              ║  ║              ║  │   ┃ built on SlateDB  ┃
╚══════════════╝  ╚══════════════╝  │   ┗━━━━━━━━━┬━━━━━━━━━┛
                                    │             │
╔══════════════╗  ╔══════════════╗  │   ╔═════════▼═════════╗
║              ║  ║              ║  │   ║  Object Storage   ║
║  Timeseries  ║  ║     Log      ║  │   ║                   ║
║              ║  ║              ║  │   ║   S3 · GCS · R2   ║
╚══════════════╝  ╚══════════════╝  ╵   ╚═══════════════════╝

OpenData Storage Engine

The shared database core behind every OpenData system. Every database compiles down to ordered SlateDB records, so every product inherits the same operating model.

One storage engine for data, indexes, compaction, and checkpoints

Scale readers independently from writers

Learn one operational model

[ learn more → ]slatedb.io →

Object Storage

The durable foundation underneath every OpenData database. Your bucket becomes the source of truth, while compute stays simple, replaceable, and easy to scale.

Store durable state in your own bucket in S3, GCS, and others

Object Storage solves replication, durability, and storage availability

Elastically scale compute without moving data

[ learn more → ]

OpenData Vector

An open-source vector database for fast ANN and full-text search on object storage.

Serve vector search from stateless nodes

Get low-latency queries at 100M-vector scale

Run production search for hundreds, not thousands, per month

[ read about Vector → ]view docs →

OpenData Buffer

A durable ingestion buffer for high-throughput data pipelines. No Kafka required.

Build HA pipelines with stateless producers and consumers

Use object storage as the durable handoff between systems

Cut broker infrastructure from your ingestion path

[ read about Buffer → ]view docs →

OpenData Timeseries

A Prometheus-compatible metrics database that keeps the architecture simple, even at high cardinality.

Ingest metrics with one writer and stateless readers

Handle millions of active series without a distributed cluster

Query with PromQL, Grafana, Prometheus, and OpenTelemetry

[ read about Timeseries → ]view docs →

OpenData Log

A durable event streaming database for millions of ordered, keyed streams. Built directly on object storage.

Route events by key, not partitions

Serve millions of keys and tens of thousands of followers from one node

Give every key its own isolated and durable stream

[ read about Log → ]view docs →

Why OpenData

Running production databases requires specialized expertise, which is why databases are expensive.

Object storage promises to change that. It gives us durable, globally addressable storage that makes radically simpler database architectures possible. What's more, the engineering needed to use object storage optimally is substantially similar across databases. That's why every OpenData database shares:

Object native

Infinite storage with high durability and excellent unit economics.

Common tooling

One operational model to learn with a common configuration.

MIT licensed

Open formats and a permissive license prevent vendor lock-in.

Read the manifesto →

Get Started

log timeseries vector buffer

$ curl -fsSL https://www.opendata.dev/install.sh | sh -s -- log

→ installed opendata-log → read the log quickstart

$ curl -fsSL https://www.opendata.dev/install.sh | sh -s -- timeseries

→ installed opendata-timeseries → read the timeseries quickstart

$ curl -fsSL https://www.opendata.dev/install.sh | sh -s -- vector

→ installed opendata-vector → read the vector quickstart

# Buffer is an embedded library — no install needed

→ read the buffer docs