Offline-first apps
Field teams, mobile apps, POS systems, and local-first SaaS keep working with zero perceived latency.
- Local writes
- Automatic sync
- Conflict resolution
SQLite, everywhere
From Edge to Cloud.
Run SQLite where your data is created: on devices, in browsers, on robots, inside agents, and across distributed systems. Sync with SQLite Cloud, PostgreSQL, or Supabase when the network is ready.
- Devices
- Browsers
- Robots
- AI agents
- Cloud systems
Edge
SQLite
Cloud
Sync
-- local SQLite, offline-first
SELECT cloudsync_init('tasks');
INSERT INTO tasks
VALUES (cloudsync_uuid(), 'inspect site', 0);
-- sync when ready
SELECT cloudsync_network_init('managed-db-id');
SELECT cloudsync_network_sync();Zero latency UX
Writes happen locally first.
Offline reliability
Work continues without a connection.
Central truth
Cloud coordinates and analyzes.
Modern applications do not live in one place anymore.
They need local decisions, offline autonomy, and a central source of truth. SQLite AI turns SQLite into the common data layer across edge devices, AI systems, and cloud infrastructure.
Product
Run local. Sync global.
SQLite Sync
Turn local SQLite databases into conflict-free replicas. Write locally, keep working offline, and sync with SQLite Cloud, PostgreSQL, or Supabase.
Explore Sync
SQLite Cloud
Use a fully SQLite-native backend for coordination, analytics, APIs, access control, backups, and multi-region scale.
Explore Cloud
SQLite Extensions
Add AI inference, vector search, agent memory, MCP tools, and JavaScript functions directly inside SQLite.
See Extensions
Use cases
One SQLite architecture for many runtimes.
AI agents with memory
Give every agent a local SQLite database for notes, state, embeddings, and task memory that can sync across workers.
- Queryable memory
- Shared knowledge
- Local autonomy
Private edge AI
Run inference, embeddings, and vector search where data already lives, then sync only what coordination requires.
- On-device RAG
- Private by design
- Low latency
Robotics and IoT
Make local decisions from sensor data, survive weak networks, and sync telemetry back to the source of truth.
- Edge decisions
- Offline resilience
- Fleet state
Postgres and Supabase teams
Keep your central system and add SQLite at the edge. Sync the local data each user, device, or agent needs.
- Bring your backend
- SQLite on the client
- Central truth
Architecture
One SQLite layer across edge, AI, and cloud.
The same embedded database can be local memory, an AI runtime, an offline-first replica, and the interface to a coordinated backend.
Local decisions
Put reads, writes, vectors, and memory next to the user or device.
Offline autonomy
Keep the product useful when networks are slow, weak, or gone.
Private edge AI
Run intelligence where sensitive data already lives.
Cloud coordination
Sync the state that needs sharing, analytics, and governance.
Ecosystem
SQLite becomes the runtime for intelligent data.
Build powerful, intelligent applications locally using our suite of extensions, then instantly scale to production infrastructure with Sync and Cloud
SQLite-Sync
Offline-first CRDT sync for local SQLite databases.
SQLite-AI
On-device model interaction and embedding generation.
SQLite-Vector
Fast vector search in ordinary SQLite tables.
SQLite-Memory
Markdown-based memory with semantic search for agents.
SQLite-MCP
Call MCP tools from SQLite queries.
SQLite-JS
Custom SQLite functions in JavaScript.
Why SQLite AI
Designed to fit your architecture, not replace it.
Not another database.
Extend real SQLite and keep its simplicity, portability, and local file-based workflow.
Not just a sync SDK.
Use CRDT sync, server-enforced access control, a managed SQLite cloud backend, and production observability.
Not a forced migration.
Start with SQLite Cloud, or keep PostgreSQL and Supabase as the central system of record.
Start building
Try the path users can feel immediately.
Create a cloud project, enable sync on one table, and watch local SQLite become a distributed application runtime.
.load ./cloudsync
CREATE TABLE tasks (
id TEXT PRIMARY KEY,
title TEXT NOT NULL,
done INTEGER NOT NULL DEFAULT 0
);
SELECT cloudsync_init('tasks');
SELECT cloudsync_network_sync();