Docs are not available for mobile use. Please use a desktop computer to view the documentation.

Getting Started - Main

Welcome to the Jaces documentation. Find guides, reference materials and examples to help you self-host or build alongside our platform.

🚨 Alert: Jaces is building towards a beta. All of the documentation herein is subject to sweeping changes. We use the docs for AI context, but feel free to poke around at your own discretion.

Getting Started

Welcome to Jaces, the personal data ecosystem that fuels AI alignment.

See the alert above: our docs are not meant for public consumption at this point. We are currently building towards a beta release. Feel free to join our Slack or reach out if you are interested in contributing.

Quick Start

Get Jaces running in under 2 minutes. The following commands only start the main services, including the web app, not the integrations— to see those visit getting started - apps instead.

# Clone the repository
git clone https://github.com/jaces/jaces.git
cd jaces

# Start all services, except for database browser
make dev

# Or start with database browser
make dev-studio

After running make dev-studio, you’ll have access to:

Service	Port	Description
Web App	localhost:3000	SvelteKit frontend - your main interface
PostgreSQL	localhost:5432	Database with pgvector & PostGIS extensions
MinIO	localhost:9000	S3-compatible object storage for files
MinIO Console	localhost:9001	Web UI for browsing stored files
Redis	localhost:6379	Task queue and caching
Processing API	localhost:8001	Boundary detection and AI processing
Drizzle Studio	localhost:4983	Database browser and query tool

How It Works

Jaces follows a “Hot & Cold” architecture inspired by how human memory works:

☀️ Hot Layer (Real-time)

What: Captures and indexes your digital activities as they happen
How: GPU-accelerated vector indexes for instant AI queries
Purpose: Answer “What’s happening right now?”

🌙 Cold Layer (Daily Batch Processing)

What: Transforms raw data into structured, meaningful memories
How: Nightly batch processing during off-hours
Purpose: Answer “What did today mean?”

Core Philosophies

Your data stays yours: Everything runs locally in Docker
No cloud dependencies: Auth proxies never store credentials
Configuration-driven: Add new data sources without code changes

Next Steps

Explore the Web UI: Connect your first data source at http://localhost:3000
Read Architecture Docs: Check /CLAUDE.md files in each directory
Join the Community: Links in the main README.md
Contribute: Review code and suggest improvements

Remember: Jaces is designed to be simple to run but sophisticated in capability. Everything you need is in Docker, commanded through Make. Your data never leaves your machine unless you explicitly configure it to.

Happy building!