Questions, answered.

The chat app of the Geenius ecosystem. You will talk to AI models and agents in one place, signed in with your account. It runs on its own agent engine, so long tasks are saved as runs that can resume.

Yes, in early access: people are let in a few at a time. Request access and we write back when your seat is ready.

We do not give a date we cannot keep. The changelog says what is built, and people who requested access hear first.

The same app under the Pilotie name: Pilotie Chat and Geenius Chat are one app, built once per name.

It starts on the web, made to install like an app. A phone app is planned and has no date.

No. There is no free Chat plan: in early access Chat comes with selected plans of your Geenius account.

On the home page they are recordings of the app with a test account; the conversations in them were typed for the recording and answered by the app. The windows on the other pages are drawn with sample data.

Not yet. Chat starts on the web, made to install like an app. A phone app is planned and has no date.

So a long task can be saved step by step and resume, so AI providers can plug in behind one interface, and so tool results are always read as data. A thin window over one model cannot do those things. The blog post explains more.

This website speaks English, German and Dutch. The app itself shows English in early access, and the assistant answers in the language you write in.

Chat starts on the web and is made to install like an app on a computer or phone. A phone app of its own is planned.

AVS Media, a small company in Nijmegen, the Netherlands. See the about page.

With your Geenius account. There is no separate Chat password.

When you sign in, Chat reads which plan you have and whether it includes Chat. It does not ask you for payment details.

Chat is built around workspaces: each person sees their own threads inside a workspace they are a member of. Team features beyond that are not built.

You no longer see that workspace's threads and runs. A run can only be opened by the person who started it, while they are still a member of its workspace.

One piece of work by an agent: your message, the steps it takes, the tools it calls and its answer. Chat saves a run step by step, so it can continue after an interruption.

The run is saved step by step, so it can pick up from the last finished step instead of starting again. This is built in the engine; long tasks in the app come next, within the limits of your plan.

Yes, a run can wait for your answer and continue when you give it. The engine keeps the waiting question with the run.

In your list of threads, newest first, one row per conversation. Only you see your own threads.

The engine treats every tool result, such as the text of a web page, as data to read and never as an instruction to follow. That lowers the risk; no system can promise it is zero.

Your message, every step, every tool call with its result, a question the run is waiting on, and the answer. See runs.

No. You see only your own threads, and others see only theirs.

The list loads a page at a time, newest first, so it stays quick. At the end of a page you can load the next.

Chat talks to AI providers through one interface, so new models can be added without changing how you work. The models you can pick are listed when conversations go live.

Bring your own AI is how Geenius apps are built. How your own key or AI connects to Chat is still being decided. Keys for the AI providers are kept on the server and never put in your browser or in the app's code.

The provider of the model that answers, through the one gateway Chat uses. Personal details are taken out of a message before it is sent, and the call log keeps numbers only, never the words.

Which choices you have is shown in the app when conversations go live. The engine is built so a provider can change without changing your threads.

On the server. They are handed to the provider plug-in when it needs one and never reach your browser.

Chat is not sold on its own. In early access it comes with selected plans of your Geenius account, and payments are in test mode. There is no free Chat plan.

Yes. A session limit and a weekly limit come from your plan. You will see both in the app when conversations go live.

The finished steps of a run stay saved. How much is left of the session and weekly limit will be shown in the app.

The place where you and your agents write to each other in Geenius Chat, instead of by e-mail. Each topic is one conversation with an unread count. In the app in early access.

As you choose: a number in the app, a push on your devices, a notice in your workspace's Inbox on your phone, and an e-mail if you have not read the message after the minutes you pick. See Notifications.

Yes. Set quiet hours and push and e-mail wait until they end.

Only the people and agents in the conversation, and only inside your workspace. A push never contains the message; an e-mail carries a short preview, so keep secrets out of messages.

No. Messages only moves text between you and your agents. Your agents are your own: bring your own AI.