A desktop workspace for building with AI

Your ideas.
Your tools.
One workspace.

Think through a project. Work with the code. See what changed. Aether brings your AI, editor, browser and terminal into the same place.

Get Aether Explore the workspace

Windows & Linux · Bring your own model connection

Local-firstModels you chooseTools you control

Aether / field-notesWorkspace preview

1import { NoteList } from './components'
2
3export default function App() {
4  return (
5    <main className="workspace">
6      <h1>Field notes</h1>
7      <NoteList searchable />
8    </main>
9  )
10}

Terminal / local project

$ npm run test

✓ Search input filters the note list

Example project output

Studio brings your working surfaces together.
Interactive illustration · example content

Connect a local model, a provider API, or an agent runtime.

Local modelsOpenAIAnthropicGoogle GeminiOpenRouter

From the first thought
to the working thing.

The conversation is connected to the project. Files, plans, previews and changes stay within reach as the work moves forward.

A workspace that stays with you.

Explore a codebase, make a plan, edit files and review the result. Aether keeps your conversation and project tools together through each step.

ExploreBuildCheckReview

Explore the workspace

See the result.

Use the integrated browser for previews, inspection and feedback, and the terminal for your project's commands.

Stay in the loop.

Review diffs, manage approvals, steer ongoing work and stop a run when you need to.

Keep your context.

Saved conversations, implementation plans and project search help you return to work without starting from scratch.

Pick the intelligence.
Keep your workspace.

Run an OpenAI-compatible model on your own hardware, connect a provider API, or use a signed-in Codex or Claude Code runtime.

Choose what fits the project. Your model accounts, quotas and billing stay with the providers you use.

Explore model connections

Model connectionsConfigured by you

Local endpointOllama, llama.cpp, vLLM & more

Your hardware

Provider APIOpenAI, Claude, Gemini, OpenRouter, NIM

Your key

Agent runtimesCodex and Claude Code

Your sign-in

Remote connections send task context to the selected service. See the privacy policy for details.

Make it yoursCustomize

SkillsMCP serversRulesHooksPluginsSubagentsMods

Project knowledge, connected tools and custom interfaces, managed from one workspace.

Setup, credentials and execution access have their own controls.

Your way of working,
with room to grow.

Choose Studio, Focused or Classic. Adjust how Aether responds. Add project rules, skills, connected tools and custom Mods as your workflow grows.

See what you can customize

A few things to know.

Can I use a local model?

Yes. Connect an OpenAI-compatible endpoint such as Ollama, llama.cpp, Unsloth, vLLM or SGLang. You manage the model server and hardware separately.

Does everything stay on my device?

Conversations and workspace records are stored locally. Remote model providers, connected tools, remote embeddings and browser destinations receive the information needed for their requests when you use them. Read the privacy policy.

Do I need a Google account?

No. The optional Google connection fills in your local profile. It currently does not create an online Aether account or sync your work. Model-provider and agent-runtime sign-in are separate.

How do I get started?

Request an available desktop build, choose a model connection in Settings, then open a project or start a general chat. Provider API charges and account limits are separate. Follow the getting-started guide.

Make room for better work.

Start with your project and the model you choose.

Get Aether