Auto
One model.
The model you picked does the whole job. It reads files, edits them and runs commands when the task needs it.
- Everyday questions
- Fixes and small features
- Drafts and rewrites
WLN is an AI chat and coding agent for the Mac. Bring a key from OpenRouter, Anthropic or OpenAI, pay the provider directly, and keep your chats on your Mac.
Free for a month · macOS 14 or later · Apple silicon
The signup form accepts an empty email. Fix it and add a test.
Run a command? pnpm test
Connect a project folder and WLN reads and changes files inside it.
Connecting a folder writes nothing into it.
Pick one model for a quick answer, two for a second opinion, or a small team for bigger changes. You choose the models. Fast, Deep and Search modes are there too.
Auto
The model you picked does the whole job. It reads files, edits them and runs commands when the task needs it.
Fusion
Two models answer the same message. A judge model reads both, writes one answer, and notes where they agreed and where they didn’t. The judge can use tools too.
Maestro
A strong model plans and decides. A cheaper Worker does the work in a private copy of your project. WLN runs the tests and looks at the screens, and only passing work lands.
WLN edits your files, runs your project’s commands and drives a browser. It asks before risky steps, lists every change it made, and lets you undo any file.
Paste a key from OpenRouter, Anthropic or OpenAI. It’s kept in your Mac’s Keychain, and requests go straight to that provider, which bills you.
WLN reads, edits and creates files in the folder you pick and runs your tests. Choose Ask every time, Allow edits or Full auto per project, and deleting always asks. Bypass permission, where nothing asks, is there too if you want it.
The AI opens pages, reads them, clicks, types, scrolls and takes screenshots. It uses a built-in browser that needs nothing installed, or Chrome, Edge or Brave with its own WLN profile.
Add MCP connectors from a gallery, remote or local. When one stops working, WLN says what’s wrong in one sentence and offers the fix.
Pages, SVGs, documents and React components open in a side panel with Preview, Code and Versions. Restore any version.
Run preview spots a Vite, Next.js or Astro project, or a static site, starts the dev server and shows the page inside WLN.
Every request counts toward the chat’s limit, Fusion and Maestro included. Near the limit, WLN wraps up with what it finished and a card to raise it.
WLN uses prompt caching, so a new message reads the earlier chat from the cache instead of paying full price for it again.
One app for questions, writing and real work in your files, with the model you choose for each chat.
Fix a bug, add a feature or write the missing tests in a project folder. Every change shows up as a diff you can undo.
Search mode looks things up on the web before it answers. Fusion gives you two takes on the same question, merged into one.
Draft, shorten or rewrite with the model you like best. Arabic paragraphs read right to left.
Let the AI go through a website for you: open pages, fill in forms and take screenshots. It asks first before anything risky.
Point WLN at the files in a folder you choose. It searches and reads them, and can make spreadsheets and PDFs.
Ask for a page or a React component and see it run in the side panel. Keep every version and go back to any of them.
Try everything free for a month. AI usage is billed by the provider whose key you use, not by WLN.
See if WLN fits how you work.
$0for 1 month
Start free monthBest value
Everything, for a year.
$19/year
That’s $1.58 a month
One payment, no renewals.
Price not set yetonce
Try it free firstPrices in US dollars. Paid plans aren’t on sale yet. Start with the free month and we’ll say here when they are.
Sign in with Apple in the app and your free month starts. Add a key from OpenRouter, Anthropic or OpenAI, and you’re ready.
Get WLN for Mac