First decide what the image is for
An illustration can have two different jobs. It can deepen a fictional companion scene, or it can be an asset with a specific composition for a creative project. The first depends on conversational context. The second depends on how precisely you can communicate visual requirements.
FunFun’s homepage describes companions that chat and exchange pictures. Adobe’s Firefly text-to-image page documents generating from text and using reference-oriented controls to influence a visual result. Those are different starting points even when both lead to a character portrait.
We compare a non-explicit original-adult illustration. We have not generated a sample, measured face consistency or tested a paid tier.
A companion scene begins with the relationship
Imagine an original 32-year-old librarian preparing a local exhibition. In the conversation, a picture might represent the quiet room where the character is sorting photographs. Its purpose is to support the shared fiction.
The helpful brief describes the situation and what should remain recognizable. It does not need to become a commercial art-direction document unless composition is genuinely the goal. Conversely, a picture that looks appealing does not prove the system understands the exhibition deadline or remembers why the character chose a particular photograph.
For this task, keep the narrative facts in a written character brief you control. Do not make an image carry information that should remain explicit in the conversation.
An illustration brief begins with constraints
Suppose you need a landscape-format portrait with empty space on the left for a heading. The character must be on the right, wearing a navy jacket, with a muted background.
Now composition is decisive. Firefly’s documentation discusses text prompts and image references as ways to communicate visual direction. A concrete brief can separate:
- Subject: an original fictional adult librarian.
- Framing: waist-up, subject on the right.
- Setting: a reading room with an uncluttered wall.
- Style: subdued editorial photography or an illustration.
- Constraint: enough quiet space for a heading.
These are project requirements, not a prompt designed to defeat moderation. A service’s support for a picture feature does not establish that it exposes every one of these controls.
For the right-aligned layout, Firefly’s composition guide gives an actual next step: Image → Generate image, choose a model, expand Composition, upload a reference, and adjust Strength before generating. A simple sketch with a figure on the right and blank space on the left communicates layout; a style reference serves a different purpose, the image’s look. Use a model/workflow that exposes these documented controls rather than assume every partner model has the same panel.
Compare the next editing step
| Your immediate need | More relevant workflow |
|---|---|
| An image that accompanies a fictional exchange | Companion context and a coherent scene brief |
| A precise layout asset | Prompt/reference controls and composition editing |
| A different character relationship | Character instructions, not a visual reroll |
| A new camera angle | A visual-generation or editing control, not more backstory |
Keep rights separate from creative fit
FunFun’s web terms describe personal use and a license over submitted input. Do not assume that an image appearing in a private account can be reused commercially or that a provider owns every part of your original concept. Read the particular grant and restriction.
Adobe’s FAQ permits commercial projects using Firefly outputs from features without a beta label; beta outputs are also permitted unless the product says otherwise. That is a material difference for this example if the portrait is destined for a business page: FunFun’s cited terms restrict its service to personal, noncommercial use. Adobe’s permission does not clear rights in a borrowed reference or guarantee exclusive copyright in the result. Check the selected model and feature’s applicable terms.
The better choice follows the deliverable. If success means a more coherent continuing character exchange, start with the companion brief. If success means a portrait that fits a page, start with the visual constraints and a tool that documents them.