Design tomorrow’s conversation first
Many character projects begin with a portrait and stop before anyone knows what to say. A more useful starting question is: what would make you return tomorrow?
For this comparison, imagine Ren, an original fictional 40-year-old potter preparing a community workshop. The character has a practical problem: too many people have booked the beginners’ class, and there is room for only one extra table. This creates a conversation about priorities rather than a list of flattering traits.
SoulFun and FunFun AI both describe custom companions. The difference worth examining is which decision their public evidence helps you make first.
A routine-led starting point
SoulFun’s download page describes character customization, conversation and photo requests. Its subscription terms separately meter texts, voice and video minutes.
If you already expect a short conversation on the commute and an occasional call at home, those units help you plan the routine. You can decide whether a call has a useful role before treating it as a feature you must use simply because it exists.
Ren’s workshop could progress through three short sessions: decide the extra table, plan the materials, and review what went wrong. A voice interaction might be convenient while you are not looking at the screen, but convenience is a reader preference. We have not tested a voice or call.
An appearance-led starting point
The FunFun creator opens with name and appearance choices inside seven stages. Its published form also defines personality, occupation, up to three hobbies and clothing selections. Those are useful broad character choices, rather than evidence of a portrait-only product. They do not establish how a later event is stored or corrected. We read the served creator without submitting a character.
For Ren, use the available personality/occupation/hobby selections only where they match the idea; keep the precise table shortage in the first conversation. Do not assume that an occupation selector offers a particular profession that has not been verified. An original first message can state: “Ren runs a community pottery workshop. More beginners booked than expected, and only one extra table fits. Help Ren choose between a second session and a smaller demonstration.” That starts the task without pretending the service already remembers a prior workshop.
| Keep in the brief | Why it helps the next exchange |
|---|---|
| Wants newcomers to feel welcome | Explains the character’s decision |
| Asks a specific question before suggesting a solution | Gives the conversation a recognizable rhythm |
| Has too little space for one more table | Creates a solvable situation |
| Does not decide the reader’s answer | Preserves participation |
Do not pay for complexity you have not used
A large character specification is not necessarily a deeper character. Ten conflicting traits can make the opening harder to direct. Begin with one motive, one obstacle and one conversational habit.
Media can be added after that concept works. A portrait might help represent the fictional workshop; it does not automatically improve the exchange about who needs the extra space. Likewise, a call is not inherently more meaningful than a clear written question.
The decision is therefore conditional. SoulFun’s public terms are more directly useful for a known communication routine. FunFun’s observable creator is more directly useful for the initial visual setup. Neither set of pages proves superior personality consistency, and neither makes a written scene brief unnecessary.
A good first result is not the longest description. It is an opening that lets the reader make a choice and gives the character a reason to care about it.