Open Source & Self-Hosted

Libre WebUI Make whatever comes next.

Turn conversations into images, interactive artifacts, research, and agent runs. Keep the workspace self-hosted, using local models or the providers you choose.

Apache 2.0 Zero Telemetry Local & Cloud AI Self-Hosted

New in v0.17.0

Everything you are working on, one tab away.

Home greets you and remembers what you were doing. Chats and Work sessions open as tabs beside it. Press Cmd K to reach any of them without losing what you were typing.

The Libre WebUI workspace: a Home tab listing chats and Work sessions to continue, beside open Work and chat tabs.
HomeTabsCmd KAgents
  • Home

    A launcher that greets you and lists the chats and Work sessions worth continuing, with a live count of running runtimes.

  • Tabs

    Chats, Work sessions, and pages stay open beside each other and survive a reload, so a project is where you left it.

  • Command palette

    Cmd K reaches anything you have, even mid-sentence in the composer. Your draft stays exactly where it was.

  • The agent you already pay for

    A Claude Code or Codex CLI installed on the server becomes a selectable chat model. No second API key.

  • A shell in the sandbox

    Work gains a real terminal inside the same locked-down container the model works in, never a way around it.

  • Your own folder, if you ask

    A Work task can be bound to a real directory instead of a container volume. Off by default, allowlisted when on.

A thought can take any form.

Begin with words. End with an image, a working interface, a clearer answer, or an agent already building — each one a tab away, and none of it leaving your machine unless you send it.

LIBRE / Creation 01 Generative composition
Created from

“Create a living poster that changes with the light.”

Output HTML / Generative
01

Imagine.

Generate and refine images beside the conversation that asked for them.

02

Build.

Turn a sentence into HTML, SVG, JSON, or code you can actually run.

03

Explore.

Bring your own documents into the conversation and follow the idea further.

04

Direct.

Hand an agent a sandboxed workspace with files, a terminal, and a live preview.

02 Your workspace

Choose the intelligence. Keep the workspace.

Use local models for private work. Reach for a cloud provider when it earns a place in the process. Every route stays visible, and every one of them stays your decision.

Explore connections
Ollama OpenAI Anthropic Google Mistral Groq OpenRouter Hugging Face ComfyUI
01 / Local

Private when it matters.

Run Ollama on hardware you own, and a local conversation never leaves it.

02 / Optional

Connected when you choose.

Add a remote provider for the work that needs one — never as a silent default.

03 / Extensible

Open by design.

Describe any OpenAI-compatible service in JSON, or talk to a coding agent already installed on the machine.

03 Begin

Start with one line.

No account to create. No workspace to rent. No trial to start. Run one command and the workspace is already yours.

npx (One Command) NODE 22.22+
npx libre-webui
Opens at localhost:8080 Install Ollama for local models, or connect the providers you choose.
01

Instant

The shortest path from a terminal to a working workspace.

02

Docker

A durable self-hosted install for a server you keep running.

03

Desktop

A native app for macOS. Windows and Linux builds are available on GitHub.

04 Extend

The workspace can grow with the work.

Connect a provider, a local endpoint, or a service you built yourself. The JSON plugin layer holds the door open without taking over the room.

my-studio.json Plugin
{
  "id": "my-studio",
  "type": "multi",
  "capabilities": [
    "chat", "image", "voice"
  ],
  "endpoint":
    "http://localhost:8000/v1"
}
Connected localhost:8000
Libre WebUI

Start in one line.Leave whenever you want.

No account, no seat, no trial. Run it on hardware you control — and if this project ever stops serving you, your installation and your fork do not need our permission.

Apache 2.0 Zero Telemetry Self-Hosted