Memory
Nothing is remembered without a review.
Memory candidates are visible before they shape a reply. Sensitive categories stay off until you turn them on. You can edit, disable, or delete any memory.
A private AI companion
Mocha keeps the things you mean to come back to — a name, a decision, a half-made plan — in one private workspace. It remembers deliberately, and it asks before it acts.
Nothing has to be tidy to be kept.
Prospect Heights · 4:42 pm
08:12 → 21:06
08:12A thought
A name. A decision. The thing due Friday. They arrive as separate pieces, each asking to be held at once. Mocha gives them somewhere to sit without asking you to organize them first.
13:40A thread
The pricing question finds the project it belongs to. The note from a walk becomes a memory you can review. A reminder takes its place in the week.
18:05A next step
Not a dashboard of everything. A small, clear view of what is open, what is waiting, and what happens next.
21:06Before it leaves
The provider, the recent conversation, the approved memories, the project brief, the connected services. Each one can be excluded before you send. Local only, ask first, or cloud assisted.
The small print
Mocha is not a place that knows everything about you. It knows what you choose to keep, and it shows you where it came from.
Your data stays in your configured database. The context preview shows what leaves.
Memory
Memory candidates are visible before they shape a reply. Sensitive categories stay off until you turn them on. You can edit, disable, or delete any memory.
Context
Local only, ask first, or cloud assisted. Before a cloud message, the context preview shows the provider, destination, memories, project brief, and connected services.
Actions
External actions are one-shot approvals. The model can propose an action, but it cannot approve or send its own request. A read grant never authorizes a write.
Start with the thing on your mind. Mocha will keep the context, and you can decide what stays.