Character chat platforms raise a consistent set of beginner questions, because several unfamiliar ideas arrive at once: characters made by strangers, personas, models, API keys, memory that fades. This page answers the questions Janitor AI characters app users ask most, in plain language and without pretending the complicated parts are simple.
Deeper guides stand behind many answers; treat this page as the index and them as the library. If your question is really a malfunction, the fixes guide is the better door than this overview.
Where an answer depends on current terms or your own configuration, the page says so. Precision about uncertainty is part of honesty.
Characters and Personas
What is a character? A created personality — premise, voice, examples — that the model performs. What is a persona? The identity you bring: who you are in the character's world. Do I need both? A persona is optional but transformative; five lines of it change how every character treats you.
Can I make my own character? Yes, and the creation guide in this library walks the whole craft, from one-line premise to example dialogue. Start through how to install Janitor AI on Android if you are new, then create early — making teaches the platform faster than browsing does.
Models and APIs
What is a model? The underlying AI that writes the replies; characters are performances, models are the performers. What is an API key? A credential that lets the app use a model under your own account with a provider — treat it like a password that can spend money.
Do I have to supply my own? No. Built-in options exist for users who prefer everything managed. The settings are an offered control, not an entrance fee, and the explainers in this library cover them when you want them.
Memory
Why did the character forget something? Because models read recent context best and distant details fade unless refreshed — the window effect. Summaries, scene-grounding, and timely reminders are the craft that carries long stories across it.
Can memory be perfect? Not with current technology, on any platform. It can be managed well enough that stories run for months, which is the honest promise, and the memory guide in this library is devoted to exactly that management. Long-story technique begins in Janitor AI chat memory and context tips.
Privacy and Safety
Are my chats private? They are processed by the platform and model to generate replies, under those providers' policies — read them once and write accordingly, keeping real personal details out of fiction. Can I block people and report content? Yes; use both without hesitation when needed.
Is it for adults? The platform serves an adult audience with content controls you should set to your own comfort. Younger users and shared devices deserve a deliberate configuration conversation, not a default. You can begin via the Janitor AI download page once you are ready to set yours.
Costs
What does it cost? The platform's own pricing and any model-provider costs are separate layers; built-in routes may be free or tiered, and self-supplied premium models bill by usage. Check current terms rather than any article's memory, including this page's deliberate generality.
How do I avoid surprise charges? Use built-in options, or set provider spending limits and check usage after your first long session. Surprises are almost always unmonitored enthusiasm, and monitoring is a five-minute habit.
When Chats Misbehave
Why did I get an empty reply? Possibly a filter, a limit, or a failed request — retry once, then shorten and simplify. Why did generation get stuck? Usually the connection or the service's moment; wait, retry, restart, in that order.
Why does my character feel different lately? Models and platforms update; also check that your persona is the active one and your settings have not drifted. The errors guide in this library decodes the common cases one by one, calmly and in order.
Characters are performances, models are performers, personas are your casting, and memory is a window you learn to manage. Hold those four ideas and most of the platform's mysteries dissolve on contact.
Your own questions, written down as they arise, will become the most useful FAQ of all — the one indexed to your stories.
When a question of yours is missing here, apply the page's four ideas to it: performance, performer, casting, window. Most puzzles on this platform are one of those four wearing a disguise, and naming the disguise is usually the moment the answer becomes obvious.
Bookmark the guides your own questions keep pointing to, in the order you actually need them. A personal shelf — memory, errors, settings, safety — organised by your history with the app will serve you better than any curriculum, because it is indexed to the stories you are really telling.
Ask your own sharpest question next — the one this page did not cover — inside the app, in a real chat, where the answer can be tested against a living conversation. That is the platform's deepest FAQ, and it is open every hour of the day.
If this page started more questions than it settled, it did its job. Curiosity with a map is the beginning of every good story told on this platform, including yours, starting tonight.
