The home page has the highlights. This is the long version — every part of the app, what it's for, and why it works the way it does. Nothing here is padding: if it's on this page, it's in the app.
Download for MacHosted for you, connected to your own keys, or running entirely on your Mac — and you can mix all three in the same app, even the same afternoon.
Aqua serves the current line-up — Claude, GPT, Gemini, Llama, Mistral, DeepSeek, Grok, Qwen and the rest — and Eaon puts them all in a single picker. Switch in the middle of a conversation and the next reply just comes back from the model you chose.
Star the ones you actually reach for so they sit up top. Rename any model to something you'll recognise later. Hide the ones you'll never touch so the list stays short.
Point Eaon at Ollama, llama.cpp or MLX and it runs the model right on your laptop. It finds the runtimes you already have installed, and if you want a new one it'll pull it from Hugging Face or Ollama — you choose the quantisation, it grabs the right file.
Once a model's on disk there's no network involved at all. Eaon keeps it warm between messages so replies stay quick, then quietly unloads it when you step away so it isn't sitting on your memory.
Already paying OpenAI, Anthropic or Google? Add your key and Eaon talks to them directly, with no middleman.
It speaks all three wire formats — OpenAI-style, Anthropic's Messages API and Gemini — so the same window works whichever one you're on. Keys go in the macOS Keychain, and every connection has its own on/off switch.
A coarse switch per connection, right in the model picker. Quiet down Aqua and keep your own Anthropic key, or the other way round.
Flip one off and its models simply leave the list until you flip it back — local models are never affected either way.
The small stuff that adds up when the app is open from morning to night.
Put two models side by side and send the same prompt to both at the same time. It's the quickest way to work out which one's better for a particular job — or to stop guessing and just see the difference for yourself.
One shortcut opens a command bar. Jump to any conversation, switch the model, flip the theme, or drop straight into a settings page — all with the arrow keys and Enter. Once it's muscle memory you'll barely touch the sidebar.
Group related chats into folders. A project's chats stay tucked inside it instead of piling into your recent list, so a week of digging into one thing doesn't bury everything else.
Tell every model how you like your answers once, and it sticks. Prefer it blank? Leave it — then nothing is quietly shaping the replies behind your back, which is the default.
Keep the conversations you keep coming back to pinned at the top, sitting above the usual by-date list so they're one click away.
Find any chat by what's inside it, project ones included. It's the same box the command palette uses, so it's always a shortcut away.
Eaon keeps count so you don't have to guess where the tokens went.
Every message is tallied — tokens in and out, a running cost estimate per model, and how your usage trends over time. Hosted models get a real number; local ones cost nothing to run, and it says so plainly.
A small gauge shows how much of a model's context window a conversation is using, live. For a local model it reads the real limit off the running server; for the rest it uses each model's published size. When it genuinely doesn't know, it shows nothing rather than a guess.
The parts that matter for privacy, spelled out instead of implied.
Every API key — Aqua's or your own — is stored in the macOS Keychain, the same place the rest of your passwords live. Never a plain-text file, never our servers.
Conversations are saved as files on your disk. Export any of them to Markdown or JSON, import them back later, or wipe the lot from Privacy settings. It's yours to move or delete.
With a local model loaded, nothing leaves the Mac — not the prompt, not the reply. Close the Wi-Fi and keep going; it won't notice.
Built in SwiftUI, not a website living inside a frame.
A real macOS app written in SwiftUI. It opens fast, follows your system, and doesn't drag a whole browser engine along with it.
Monochrome by default, following your system theme, with a few accent options if you want a little colour. Both themes are designed together, so neither one feels like an afterthought.
IBM Plex throughout — the monospace for labels, model names and code, the sans for actual reading. Same family, so the whole thing reads as one piece.
When there's a new version, a small card slides in. One click downloads it — you won't have to come back to this site to stay current.
First launch walks you through connecting a model and finding your way around, then gets out of the way and lets you get on with it.
Keyboard navigation everywhere, tidy empty states, sensible defaults, and copy that tells you what's actually happening. The bits you only notice when they're missing.
Free forever, and open source. Your keys, your models, your history — kept on your machine.
Download for macOSUniversal · macOS 14 Sonoma or later · v2026.1.8