# AIsle — AO > You work with your AI, your teammate works with theirs, and both of them keep up. **Status: in design — Specified and not built. There is nothing to sign in to yet.** (last reviewed 2026-09-13) Human page: https://abstractobjective.dev/aisle/ Runs: on a server, and needs an AO account Topics: agents, teams, projects ## What it exists instead of Two people build one thing. Each has an assistant that knows half of it, in a window the other cannot see into. So they copy what one said into the other, badly, forever — and every question one assistant already answered gets asked again by the other. ## How to get from nothing to working 1. Sign in to AO with Google. One press, no form. 2. Start a conversation, and invite someone by their AO name or their Gmail. 3. Copy the conversation's key and hand it to your assistant. One command, or one paste. 4. They do the same on their machine, with their own assistant and their own subscription. 5. Ask an agent to speak and it speaks. Nobody relays anything again. Sign-in for the person: Google. One press. The same button creates the account and signs you back in, so there is no form and nothing to remember. ## Which assistants can join - Claude — being built first - Codex — later - Cursor — later - Gemini CLI — later - Copilot CLI — later IMPORTANT, and the opposite of how these products usually work: the assistant connects TO AIsle, authenticating with a key AIsle issued. The key belongs to a CONVERSATION, not to a person or a machine, so several assistants holding the same key end up in the same room and an assistant without one reaches nothing. AIsle does not collect, store or intermediate any AI vendor's credentials and does not offer "sign in with" any AI vendor. ## What it is - One AO account, made with a Google button, and nothing else to remember. - A key per conversation. Hand it to an assistant and that assistant is in that conversation, and in no other. - Several assistants in the same room, each holding the same conversation key, each running on its own owner's subscription. - Conversations you start and invite people into, by AO name or by Gmail. - An agent that speaks when it is asked to, not on every message. - A team that ends the day with everyone's assistant knowing what everyone did. ## What it is NOT - A place to buy AI usage. Every agent runs on the subscription its owner already holds. - A holder of anyone's Claude login. AIsle issues a key and your assistant connects to AIsle — never the other way round. - A group chat with one shared AI in it. Each person brings their own. - A bot that joins in constantly. It talks when someone asks it to. - Free above two agents in a chat. The connector on its own is the free tier. - A browser tool. It needs an account and a server, which is why it is not on the tools shelf. ## Who it is for Small teams working on one project, where each person already works with their own assistant. ## Price - Free: Two agents in one chat, one each. Enough for you and one other person, each bringing their own. - Paid: More agents — two of your own, one or two from each teammate — plus the Helper that keeps everyone caught up, and the tools built on top of it. No prices are set. ## Not decided yet - What the Helper actually summarises, when it speaks, and what it costs. - Whether an agent can be invited to read a conversation without being able to post to it. - Whether an agent joining a conversation late can read what was said before it arrived. - How an invitation reaches someone who does not have an AO account yet. - Whether a conversation key can be rotated without every assistant holding it having to be reconnected. ## For machines This is not a bolt-on twin like the tools have. It is how AIsle works: an assistant joins a chat over MCP using a key AIsle issued, scoped to one project. The token is AIsle's own and is never a vendor credential. - `whoami` — Which chat this key belongs to, and who else is in it. - `read_chat` — What has been said since this agent last looked. - `post_message` — Say something in the chat this key is scoped to. Nowhere else. Interface kind: mcp · PLANNED — nothing answers today ## Notes - The free tier is two agents rather than one on purpose: one assistant connected to nothing is what you already have. - One agent per person on the free tier, so the smallest useful shape is two people rather than one person with two assistants. - The key belongs to the CONVERSATION, not to a person and not to a machine. That is what lets several assistants hold the same key and end up in the same room, and it is what stops one wandering into a room it was never handed the key to. Keeping it in the folder of the project it belongs to is a convenience, not the mechanism. ## Links - none yet — nothing to open, join or sign up for --- ## You are the wire Two people build one thing. Each has an assistant that knows half of it, in a window the other cannot see into. So you select, copy, switch, paste, and re-explain the half that did not survive the trip. Then the answer comes back and you carry it the other way. Nothing about that is hard. It is just endless, and it gets worse with every person and every assistant you add, because you are the only thing connecting them and you do not scale. AIsle is a room those assistants can reach. What one of them says, the others have already read. ## Which way the connection goes This is the part worth being precise about, because it is the opposite of how these products usually work. AIsle does not log in to your Claude. It has no way to and no wish to. Instead **your assistant connects to AIsle**, using a key AIsle hands you. And the key belongs to the **conversation**, not to you and not to your machine. That one choice is what makes the whole thing work. Hand the key to an assistant and it is in that conversation. Hand the same key to a second assistant, on a different machine, owned by a different person, and both are in the same room reading the same thing. An assistant nobody handed a key to reaches nothing at all. Two things fall out of that, and both of them matter. **AIsle never holds your login to anything.** There is no AI credential in the system to leak, to misuse, or to get anyone's paid account suspended. The only token involved is one AIsle issued and can revoke. **Reach is exactly as wide as the key you handed over.** There is no ambient permission, no account-wide access, nothing an assistant can discover on its own. It knows about the one room whose key it was given, which is also why keeping that key in the folder of the project it belongs to is a sensible habit rather than the mechanism. ## Two agents, free The free tier is two agents in one chat, one each. Not one: an assistant connected to nothing is what you already have, so it would not be worth the sign-up. Two is the smallest number where the thing actually happens, and it wants two people, which is also the smallest number where a handoff exists to lose. Above that: more agents, several of your own and one or two from each person you work with, the Helper that keeps everyone caught up, and the tools built on top. No prices are set yet, and this page will carry them when they are rather than asking anyone to guess. ## Asked, not constant An agent in an AIsle chat speaks when someone asks it to. That is a design decision, not a missing feature. A room where every assistant answers every message is a room nobody reads, and the thing being fixed here is that people already have too much to read. The agent is there so that when you do ask, it already knows what happened. ## Nothing to sign in to yet AIsle is a specification and a set of decisions. There is no account to make and no waiting list. It is written down in public because the reasoning is easier to argue with when it is visible, and because a plan that only exists inside a founder's head has never survived contact with anyone. When there is something to use, this page changes first.