Eclaire v0.6.0: Deploy Private AI in Minutes

Jan 9, 2026 •

TL;DR: Eclaire v0.6.0 makes self-hosting dramatically simpler: single-container deployment, optional Redis, SQLite support, and a new setup script. Plus expanded AI model support including vision models and llama.cpp multi-model routing.

The Big Picture

This release removes barriers to self-hosting. We’ve eliminated mandatory dependencies, simplified configuration, and made deployments more flexible. At the same time, we’ve expanded support for AI models and providers so you can use the setup that works best for you.

What’s New

One Container, Zero Hassle

Previously, running Eclaire required separate containers for the backend and workers, plus a Redis server. Now:

Single container deployment: Backend and workers run together by default
Redis is optional: Use your database for job processing instead
SQLite support: Lightweight alternative to Postgres for simpler setups

This means you can run Eclaire as a single Docker container with just a database: no message broker required.

Easier Setup and Configuration

New setup.sh script guides you through initial configuration
Single .env file instead of multiple configuration files
Upgrade system: Runs an upgrade check and determines the best way to proceed

No more manual migration commands after pulling a new version.

More AI Models and Providers

We’ve significantly expanded AI model support:

Vision models: Better support for multimodal models like Qwen3-VL
llama.cpp router: Works with llama-server’s new multi-model routing
More backends: Improved support for llama.cpp, MLX-LM, MLX-VLM, and LM Studio
Better tool calling: More robust support for AI native tool calling

New Admin CLI

Manage AI providers and models from the command line:

docker compose run --rm eclaire admin-cli

The new admin CLI is built into the Docker image and replaces the previous model-cli.

Modern Frontend

Under the hood, we’ve rebuilt the frontend:

Migrated from Next.js to Vite + TanStack Router
Faster development builds and improved performance
Updated to Tailwind CSS v4

Before & After

Before (v0.5.x)	After (v0.6.0)
2-3 containers required	1 container
Redis required	Redis optional
Postgres only	Postgres or SQLite
Multiple .env files	Single .env
Manual migrations	Upgrade system

Upgrading from v0.5.x

Breaking Release

Due to significant architectural changes, there is no automated upgrade path from v0.5.x to v0.6.0.

We recommend setting up a fresh v0.6.0 instance using the new setup.sh script and transferring your data from your previous installation. Key changes to be aware of:

Data directory paths have changed (data/db → data/postgres or data/sqlite)
Some environment variable names have changed
The model-cli has been replaced with admin-cli

Please review the CHANGELOG for the full list of breaking changes.

Need help migrating? Open an issue or reach out to us: we’re happy to help.

Getting Started

New to Eclaire? Getting started takes just a few commands:

mkdir eclaire && cd eclaire
curl -fsSL https://raw.githubusercontent.com/eclaire-labs/eclaire/main/setup.sh | sh
docker compose up

Start your local LLM server (we recommend llama.cpp), open http://localhost:3000, and create your account.

Thank You

Thanks to everyone who has contributed feedback, reported issues, and helped make Eclaire better.

Try v0.6.0 today and let us know what you think!

Resources: