PACIFICDB COMMUNITY v1.2.0-rc.2 · PRERELEASE

Built for data
that refuses
one shape.

Documents, vector search, resumable media, verified backups, and RF3 replication—unified in one self-hosted database you control.

Download PacificDB ↓See how it works →

Local-first · Open source · No hosted service required · v1.2 changes and evaluation guide →

DocumentsCRUD · filters · aggregation
VectorsSimilarity + metadata filters
MediaResumable · checksummed chunks
QueriesDocuments · vectors · media
PacificDBone engine
WAL
LSM
Indexes
Snapshot
RF3
Data does not stay in one boxGive every shape one durable home.
QueryJSON documents
SearchIndexed vectors
ProtectVerified backups
AvailableRF3 Raft

HOW PACIFICDB WORKS

Every shape enters.
One engine carries it.

JSON records, embeddings, and file chunks move through a shared engine for querying, indexing, durability, recovery, and replication.

Automatic system walkthrough
01 · Ingest
DocumentJSON record

Filter, project, paginate, aggregate.

{ }
VectorEmbedding

Similarity search with metadata filters.

↗
MediaFile chunks

Resumable, sequential, SHA-256 verified.

≋
PacificDB enginelocal / self-hosted
QueryCRUD + explain
VectorSimilarity index
MediaChunk verifier
IndexesB-tree + vector
DurabilityWAL + LSM
RecoverySnapshots
RBAC / TLSRF3 RaftBackup / restore
Structured queryMatching documents

Return projected, bounded JSON results.

✓
SimilarityNearest vectors

Rank matches using supported metrics.

≈
Media retrievalVerified content

Resume or download stored file chunks.

↓
INGEST

Applications send JSON documents, vector embeddings, or resumable media chunks through the same PacificDB deployment.

ProtectVerified backup & restore

Create snapshots, verify them, export every physical file, and inspect restore history.

SecureTLS, keys & boundaries

TLS/mTLS, passwords, API keys, RBAC, tenant isolation, and audit logs.

ObserveMetrics & diagnostics

Health checks, Prometheus metrics, logs, and storage diagnostics for operators.

ScaleRF3 & explicit sharding

Raft elections and catch-up plus explicit placement, migration, and rebalance controls.

WHY PACIFICDB

One system of record.
More room to build.

PacificDB keeps related application data together without forcing you to operate a document database, vector store, media pipeline, and backup path as separate products.

011×

One path for every shape.

Write documents, embeddings, and resumable media through one database, one protocol, and one operational model.

02W

Recovery is engineered in.

WAL, LSM storage, compaction, checksums, snapshots, and crash recovery protect more than the happy path.

03R3

Distribution stays explicit.

RF3 Raft handles elections and follower catch-up; sharding controls placement, migration, and rebalance.

04⌂

Run where your data belongs.

Native packages, Docker, Kubernetes, Helm, CLI tools, and three SDKs carry you from local work to controlled deployments.

BUILT BEYOND ONE MODEL

Documents are the start.
Not the boundary.

Use the model that fits each part of your application while keeping the data close to the same durable core.

Query the shape you know.

Store JSON documents with filters, projection, pagination, bounded aggregation, explain metadata, and secondary indexes.

DOCUMENT WORKFLOWS INCLUDED

Search by meaning.

Store finite numeric embeddings, rank nearest matches with vector indexes, and narrow results with metadata filters.

SIMILARITY WITHOUT A SECOND STORE

Move files without losing state.

Upload images, GIFs, audio, video, and arbitrary files in bounded chunks with SHA-256 verification and explicit resume IDs.

MEDIA THAT CAN PICK UP AGAIN

Under every model: WAL, LSM storage, checksums, snapshots, recovery, RF3 Raft, and operator-controlled backups.

Explore the exact scope ↗

PACIFICDB WORKBENCH · v1.2.0-rc.2

Explore your data
from your desktop.

Work with documents, vectors, and media in a desktop app that starts its own PacificDB engine. Your data stays on your computer and persists between sessions.

Linux x64 .deb ↓Windows x64 .exe ↓macOS Apple Silicon ↓macOS Intel ↓Install guide →

Pre-release v1.2 · RC.2 · Ubuntu 24.04 x64, Windows x64, macOS 15 Intel/Apple Silicon · SHA-256 checksums. Signing/notarization and enterprise endpoint policies are not fully qualified; your IT team may need to approve installation.

Direct database connections, generated SDK examples, and bounded Overview count loading. See what changed.

PacificDB WorkbenchDesktop app with bundled engine and CLI
  • Manage local dataOpen databases directly and organize collections without mandatory project selection.
  • Run queriesInspect documents and search vector data in one workspace.
  • Resume where you left offLocal database files persist when the app closes.

INSTALL ON UBUNTU 24.04

sudo apt install ./PacificDB-Workbench-1.2.0-rc.2-linux-amd64.deb

RUN IT WHERE DATA LIVES

Your infrastructure.
Your database.

Start with an installer that includes the engine and native CLI. Docker, Kubernetes, and Helm deployment files are available in the repository.

v1.2.0-rc.2 PRERELEASE Pre-release v1.2 · Company testing · Release notes and qualification limits →

L_

Linux

Ubuntu 24.04 · x86-64

.deb package · v1.2.0-rc.2 prereleaseDownload deb ↓
  1. Download and verify the package.
  2. Install sudo apt install ./pacificdb-community-1.2.0-rc.2-linux-x64.deb
  3. Open pacificdb; the local engine starts automatically.
⊞

Windows

Windows · x64

Workbench installer · bundled engine and CLIDownload Windows ↓

Company-testing prerelease. Install and open PacificDB Workbench.

⌘

macOS

DMG · bundled engine and CLIDownload macOS Apple Silicon ↓

Verified on macOS 15. Signing/notarization policy is not fully qualified.

Pre-release v1.2 is for evaluation and is not certified production-ready. Four desktop packages are published and install-tested. Verify checksums and use isolated test data. npm client/CLI are available; Python wheel and Java JAR/POM are on GitHub while registry setup remains pending. Read the qualification limits.

Download checksums ↓Build from source ↗

FROM ZERO TO YOUR FIRST WRITE

Start local.
Build outward.

Install the engine, open the shell, and connect an application without handing your data to a hosted service.

Read the documentation →
01

Start one local engine

After installing PacificDB 1.2.0-rc.2, launch the loopback engine and enter the interactive shell.

pacificdb

Local mode binds to 127.0.0.1:9000. Configure TLS and authentication before any network deployment.

02

Shape your first data

Create a database, then a collection. The updated source CLI selects a created database automatically:

create database hello\ncreate collection ideas\ninsert ideas {"id":"first","title":"Hello PacificDB"}\nfind ideas {"id":"first"}
03

Connect the application

Install the 1.2.0-rc.2 Node.js client and connect it to the same local engine.

npm install @pacificdb/client@1.2.0-rc.2

Save the example below as hello.mjs, then run node hello.mjs.

ONE PROTOCOL. THREE CLIENTS.

Meet your data
where your code lives.

Connect through Node.js, Python, or Java. This first write uses the PacificDB v1.2.0-rc.2 Node.js client and the Community JSON-over-TCP protocol.

Node.js 1.2.0-rc.2 package ↓CLI 1.2.0-rc.2 package ↓Python install guide →Java install guide →
hello.mjs
import { PacificDB } from '@pacificdb/client';

const db = await PacificDB.connect('pacificdb://127.0.0.1:9000/app');
try {
  await db.createDatabase(); // Omit this if the URL database already exists.
  await db.createCollection('events');
  await db.insert('events', { id: 'event-1', type: 'signup' });
  console.log(await db.find('events', { type: 'signup' }));
} finally {
  db.close();
}
Local connection · JSON over TCP

OPEN SOURCE. SELF-HOSTED.

Give every shape a home.

Explore PacificDB Community v1.2.0-rc.2 and help build a database that keeps more of your stack together.

Explore PacificDB on GitHub ↗