Personal apps that remember you
An app that remembers you can carry useful context into your next conversation: who you cook for, what you're working toward, or why you started a list. In Kaddy, that memory belongs to the individual app, and you can read and edit it.
Saved data, chat history, and memory do different jobs
Your meal plan is data: Monday's dinner, the ingredients, and any notes you've saved. Chat history is the conversation where you discussed the week. Memory is the useful context the assistant carries forward, such as “usually cooks for two.”
Those aren't interchangeable. If a detail must be recorded exactly, put it in the app's records. Don't assume a remembered summary is a complete or permanent copy of everything you've said.
A small example: dinner on a busy week
Say you've told the meal planner you cook for two and prefer quick weeknight meals. When you return to discuss another week, that context can save you from explaining the same setup again.
Now suppose a friend is staying for a month. Tell the assistant what changed, and inspect its memory if the old assumption keeps showing up. “Cooks for two” was useful until it wasn't. Editable memory matters because ordinary life changes.
This is an example of how you might use an app, not a guarantee that every suggestion will be correct. Check the plan before relying on it.
Each app remembers its own conversations
Kaddy separates the conversation you use the app for from the conversation that builds it. The regular chat is the source for memory reflection; the build log isn't quietly turned into your personal profile.
Your other apps don't automatically receive that memory. A friends tracker doesn't know your dinner plans just because both live in Kaddy. That keeps each assistant's context tied to its job.
Keep the memory worth keeping
Open the app's memory view and read what's there. Correct a preference that was misunderstood. Remove context that no longer belongs. When a fact needs an exact date, number, or status, use a field in the app rather than relying only on prose memory.
Also check the selected model. Some Kaddy models run on Cloudflare and others use an external provider; editable memory doesn't mean inference always stays on one host.
Useful memory saves repetition
The point is fairly ordinary: fewer introductions to an app you've been using for weeks. A reading assistant can carry context about your interests. A project assistant can keep the purpose of a project in view while you work through its details.
Start with a routine where you find yourself repeating the same background. If you mostly need a list, a list may be enough. The meal planner brief shows how structured records and a conversation can each do their part.