Introducing moccet-assistant

Every day you generate a rich trail of data across the tools you use. Messages, meetings, emails, health metrics, documents, searches. Taken together, they already contain what an AI would need to give you answers that are actually specific to you and your life.

Today we are introducing moccet-assistant. It is a personal AI that connects to the tools you already use and starts every answer from your own context.

Your whole picture, one conversation

Connect Slack, Gmail, Calendar, GitHub, Apple Health, and more. Once connected, the assistant draws from all of them simultaneously, alongside real-time web search, and gives you one clear answer.

Say you ask "catch me up on what I missed yesterday." Within seconds you get a full briefing built from your actual Slack threads, emails, and calendar. The decisions your team made, the emails that need replies, the meetings coming up, and the action items that landed on your plate.

You have a meeting with a new client tomorrow and cannot remember where things left off. The assistant pulls your last email thread with them, the notes from your previous call, their company's latest news, and your schedule. You get a briefing with the context you actually need, built in seconds from data that would have taken you thirty minutes to piece together manually.

It works for personal things too. Want a weekend away. The assistant knows your calendar is free on Saturday, finds what is available, compares options, and lays everything out. Want to eat better based on how you have been sleeping. It reads your wearable data and recent meals and gives you something specific to the past week, not a generic plan pulled from the internet.

All of this works the moment you connect your accounts.

How it works

When you ask a question, moccet-assistant decomposes it into subtasks and assigns each to the best available model and data source. One agent reads your Slack. Another scans your inbox. Another checks your calendar. Another searches the web. They run in parallel. We call this the moccet-network.

You see all of it happening. The interface shows which services are being queried, what searches are running, and which sources have been read. When the answer is ready you can expand a panel that shows the full trace. Which models were involved, how long each step took, what it cost in credits. Full transparency into every answer.

We are model-agnostic by design. The system routes across multiple frontier models and picks the best one for each subtask. Fast models for simple lookups. Reasoning models for complex synthesis. The routing adapts as models improve. You always get the best available intelligence applied to your specific question without having to think about which model to use.

Actions, not just answers

When the assistant finds something that needs doing, it gives you interactive tasks directly in the conversation. Approve a Slack reply. Reschedule a meeting. Delegate a follow-up. Send an email. Every action shows you exactly what will happen and asks for your confirmation before executing.

Some tasks need a decision first. The assistant asks you inline. "Sarah asked if the team dinner is still on for Friday. Confirm, or suggest moving to next week?" You pick. It follows through.

The view shows your top three items first. Expand the list if you want everything. Not every response generates tasks. Research questions return research. Planning questions return options. The assistant adapts the format to the question.

Built for how you actually work

Type a follow-up while the assistant is still working and it queues automatically. A count shows how many messages are waiting. Dismiss any before they are sent. When the current task finishes, queued messages process in order.

Tap the mic and talk. Words appear in the input field as you speak. No separate screen. Stop, review, edit, send. Voice and text share the same interface.

Attach files, photos, or paste links from the input bar. Point your camera at something and it becomes part of the conversation. The assistant reads PDFs, images, spreadsheets, and documents alongside everything else it knows about you.

When an answer is useful, share it directly. Copy, Slack, email, WhatsApp, Notes. One tap.

Your data stays yours

Every action the assistant takes requires your explicit confirmation. There are no background tasks and no autonomous decisions.

All data is encrypted in transit and at rest. Connected credentials are stored in a secure vault. The system never trains on your personal information. Disconnect any service, delete your history, or export your data whenever you want.

Get early access

moccet-assistant is in private development on iOS. We are building toward something we think matters. The same assistant, same memory, same understanding of your world, available on every surface you already use. Slack, WhatsApp, iMessage. Start a conversation on your phone, pick it up on your laptop.

We are opening early access to a small group of people who want to help shape what personal AI becomes.

Request access at moccet.ai or reach out at team@moccet.com.