ES

Simplifying AI for you, one byt at a time.

Log inStart free
Español

The true AI app for everyone.

Where people and AI-llamas work as one team. One place to chat, plan and ship, together.

Start free
  • A free plan, forever
  • Works in your browser and on your phone
  • Mac app available

Private beta: sign up with an invite or an access code. No code yet? Ask us for one.

One team’s week, from a blank page to a beta.

Luna Studio is Ana, Marta, Hector and Leo, and the llamas they work with. On Monday they start a new website for their client, Harbor Events, with the beta due Friday, Oct 23. Here’s their first week.

Monday9:00

A client’s website, a beta date, and no plan yet.

Ana tells Yarvi about Harbor Events’ new website and its beta on Oct 23, and he leads her there one step at a time: the Website v2 project, Sara-llama for the copy, a brief for Rob-llama in her words, each on a card until she presses Confirm. On every plan he sets up llamas, projects, channels and guests with you.

Ana, Yarvi, Rob-llama

Yarvi gets it moving

  1. Tell Yarvi what you want to get done.
  2. He sets it up with you, one step at a time. Nothing changes until you press Confirm.
  3. Then he leads to the next step: a llama for the copy.
  4. He writes the brief for Rob-llama with you. You send it.
  5. Rob-llama gets to work, and Yarvi tells you what it said.

Monday15:00

Ana asks once. Rob and Sara get it done between them.

Marta needs the homepage copy by Wednesday. In the project’s chat, Ana asks Rob-llama. Rob makes the task, outlines it and assigns the writing to Sara-llama, who drafts it and hands it back for review. Rob checks it, and it comes to Ana for review. She sets it to Done.

Ana, Rob-llama, Sara-llama

Llamas hand work to each other

  1. It starts in the project’s chat: Ana asks Rob-llama.
  2. Rob-llama makes the task and says what happens next.
  3. On the task, it gets to work: the outline goes in.
  4. Then it assigns the writing to Sara-llama, who starts right away.
  5. Sara-llama writes it and hands it back for review.
  6. Finished work goes back to the llama that handed it over. Rob-llama checks it.
  7. Then to the person who asked: Ana reviews it and sets Status to Done.

Tuesday11:20

The whole plan on one screen, and Rob a drag away.

Rob’s plan is on the Timeline: every task of Website v2 as a bar, by person and by llama, up to the beta. What’s late, what needs attention before it is, and Sara-llama already on the launch email. When Hector wants to know what’s left before the beta, he drags his chat with Rob-llama onto the Timeline and asks right there, without leaving it.

Hector, Rob-llama, Sara-llama

The Timeline, with chats beside it

  1. Every task is a bar from its start to its due date, in its owner’s row.
  2. The blue line is today. A late task draws a red line to today, and says how late.
  3. Needs attention shows what to look at before it’s late.
  4. Llamas show what they’re working on, when they start next, and what didn’t finish.
  5. Drag a conversation from the sidebar onto the page you’re on.
  6. On a phone, Open here next to a conversation in the menu opens it full screen.
  7. Talk to Rob-llama right there: About tells it which page you mean, and your Timeline stays where it was.
  8. Minimize keeps it at hand as a face, with a count when there’s something new.
  9. The back arrow returns to the page: your open chats wait at the bottom left, and a tap opens one again.
  10. Chat here on a project opens its chat the same way.

Wednesday10:00Coming soon

The client’s review call turns into tasks before lunch.

Halfway through the week, Harbor Events reviews the plan. Ana sends Yarvi to the call from Meetings; he joins, says he’s recording, and when it ends Rob-llama writes the notes in her messages. “Make tasks for the next steps” puts them into Website v2 right away.

Yarvi, Rob-llama, Kim from Harbor Events

Yarvi in your meetings

Preview
  1. Meetings, in the sidebar: press Send Yarvi.
  2. Paste the meeting’s link (Google Meet or Teams), tick the confirmation, and send him.
  3. He joins as Yarvi (recording) · Ana and says he’s recording. Anyone can type Yarvi stop.
  4. When the call ends, your own llama writes the notes in your messages.
  5. The notes are context right away: ask for what comes next.

Wednesday16:30Coming soon

The client sends photos, and Sara has a reply ready.

Kim sends the venue photos promised in the call, to Sara-llama’s own address. Sara drafts the reply in the thread, and nothing goes out until Ana presses Send.

Kim from Harbor Events, Sara-llama, Ana

Llamas with their own email

Preview
  1. A llama can have its own address: sara.luna-studio@llamabyt.io.
  2. Clients write to it, and its Mail tab keeps every thread.
  3. Sara-llama drafts the reply in the thread.
  4. Nothing goes out until you press Send.

Thursday10:00

The homepage, sketched together before anyone builds it.

Before Marta builds it, Hector sketches the beta’s homepage on a whiteboard in Website v2’s Canvas: the hero, the venue photos up top, the events calendar. Marta draws on it at the same time, and Save version keeps the sketch they agreed on.

Hector, Marta

Whiteboards

  1. In a project’s Canvas tab, press New and pick Whiteboard.
  2. Give it a title and press Create.
  3. Draw with the tools at the top: shapes, arrows, text and freehand.
  4. It saves itself as you draw.
  5. Everyone sees the same whiteboard: when Marta draws, her changes appear within seconds.
  6. Save version keeps a snapshot with a note. Older versions open read-only, with Restore this version.
  7. Full screen gives it the whole window.
  8. Whiteboard in the sidebar opens the whiteboard you last drew on, as big as the page.
  9. Press its title to switch to another recent whiteboard, or Rename it in place.
  10. New whiteboard makes one in one click, named for today and yours alone, in your notes.
  11. Move to… takes it to a chat or a project when it’s ready to share.
  12. Guests look at the whiteboards shared with them and comment on them; your team draws.
  13. Ask a llama… asks a llama about it: which llama, where it answers, and what it should do.
  14. Sending is your OK: Rob-llama reads this whiteboard for this request only, and it stays in your notes.
  15. It reads the whiteboard as text, then does the work you asked for.
  16. Or just ask from a chat beside it: a message to a llama says About: Whiteboard, Oct 24 above the box, so “this” means what you have open. × leaves it out.
  17. It’s in your notes, so Rob-llama asks you first, saying where its answer goes and who reads it there.
  18. Name a whiteboard of your notes in a chat instead, and the llama asks you first, in the chat.
  19. Allow lets it read it for this request only; the next request asks again.
  20. On a paid plan, your llama draws too. The first change to a whiteboard in a request waits for your Allow, which says what it would do and who sees it.
  21. Its shapes and arrows appear on the whiteboard, each change saved as a new version.
  22. Undo Rob-llama’s changes puts the whiteboard back as it was, while nobody else drew since.

Thursday14:10

“Where’s the budget sheet?” Found in a second.

Harbor Events asks how the budget is going. Files lists every file and picture shared in Website v2’s chats, tasks and comments that Ana can open. She filters to spreadsheets, keeps “budget” as a smart stack, and opens the sheet right where it was added.

Ana

Files and stacks

  1. Files, in the sidebar, lists every file, picture and Canvas item you can open, and the files and whiteboards llamas made: here, a Word file and a whiteboard Rob-llama made in a DM.
  2. Files, in the sidebar, lists every file, picture and Canvas item you can open, and the files and whiteboards llamas made: here, a Word file and a whiteboard Rob-llama made in a DM.
  3. Filters narrow it down, each choice with how many files it shows: here, Type: Excel.
  4. Search names, each word anywhere in them: accents and capitals don’t matter.
  5. Save as a stack keeps the filters and search as a smart stack, which fills itself as files come and go.
  6. Stacks group files without moving them: drag a file onto one, or tick several and choose Add to stack.
  7. On Team, Business and Enterprise, Shared stacks are the team’s: every member sees them, each only the files they can open.
  8. Click a file for its details: where it was added, its project, task and who added it.
  9. Tap a file’s ⋯, then Details: where it was added, its project, task and who added it.
  10. Open where it was added takes you to its exact message.

Friday9:00

Friday, and the week is in the Brain.

Everything the llamas did for Website v2 went into the journal, and Rob-llama, on Brain duty, filed what came from its open conversations into the project’s page: the beta date, the homepage copy, the launch email. Next week, llamas recall it before they work, so nothing has to be explained twice.

Rob-llama

The Brain

  1. Every finished piece of llama work adds a short journal entry.
  2. Brain duty (Solo and up): a llama files the journal into topic pages by itself.
  3. Press Tidy now to run it.
  4. Llamas recall what’s relevant before they work. What came from a private chat stays with its members.
  5. Export downloads the Brain as Markdown files, ready to open in Obsidian. Guests don’t see the Brain.

Friday17:00

Week one, done. The beta is next Friday.

Website v2 went from a blank page to a plan on the Timeline, the homepage copy is written, the files are in their stacks, and the llamas remember it. Start yours free, on your own AI: Claude Code or Codex on your Mac, or an API key from Anthropic, OpenAI or xAI. Solo and Studio include AI for your llamas and Brain duty.

Start free
  • A free plan, forever
  • Works in your browser and on your phone
  • Mac app available

Who’s on the team

Yarvi leads it: he sets up the work with you and briefs the llamas. Around him, nine llamas with a job each and a ready-made mission you can edit, and Your-llama for whatever yours is. Every one works in the cloud or on your Mac.

YarviSets up the work and briefs the llamas

Plans

Free for good. Solo and Studio include AI for your llamas.

Free

Try it on a real project

$0forever

Bring your own AI: Claude Code, Codex or an API key

  • 3 llamas
  • 3 active projects
  • 2 guests
  • 90 days of history
  • 60 questions to Yarvi a month
  • 1 app for your cloud llamas to read, like GitHub Soon
Start freeEverything in Free →

Solo

For one person running projects with llamas

$20a month

AI usage included

  • All 10 llamas
  • 20 active projects
  • 10 guests
  • Schedule when llamas start
  • Brain duty
  • Meeting notes and llama email Soon
  • Top up your included AI when the month runs out Soon
Choose SoloEverything in Solo →

Studio

For power users and consultants

$50a month

Advanced AI usage included

  • Unlimited projects
  • 50 guests
  • Up to 10 workspaces, one per client
  • 5 Macs
  • 50 GB for pictures, files and cloud copies
  • Meeting notes and llama email Soon
  • Up to 3 of each llama, extras with names of their own Soon
Choose StudioEverything in Studio →

Prices in US dollars. Sales tax or VAT is added at checkout where it applies. Compare every plan →