Create a Two-Eye Avatar: Flat Mascot or Soft-Vinyl Character
Choose the right GrokBot Avatar mode, write a precise brief and review the delivered image at profile size. Prepare offline without submitting an image-generation task.
Fastest way
Ask your AI to do it
Paste this into the AI agent you already use:
Use GrokBot Avatar to prepare one avatar brief. First choose Original flat mascot or Anime soft-vinyl character from my request. Read the selected references, preserve explicit character requirements and clarify only missing personal essentials without inferring demographics. Prepare an English prompt and a full-size, 48-pixel and circular-crop review plan. Do not call image generation, a paid API or a fallback service. Say clearly that a brief is not a generated image.
What does GrokBot Avatar make?
GrokBot Avatar v1.0.1 is an image-generation instruction package with two distinct directions: Original makes a flat geometric mascot; Anime makes a rounded soft-vinyl toy character. Both use two simple solid eyes. It defaults to one square PNG requested at 1024 × 1024 pixels. Requested dimensions are not proof of delivered dimensions: inspect the actual file. This package does not include a local generator, SVG editor or animation exporter.
The free download requires a verified Skillrove account. This release specifies the host’s built-in image_gen tool and forbids a FAL or CLI/API fallback. Check whether your agent actually exposes that tool before planning a generation. Model access, usage limits and possible charges belong to the host service; a free package does not make generation free.
Choose a silhouette or a character before writing the prompt
If the request is only “an avatar”, the package defaults to Original. A person, character portrait or humanized avatar selects Anime. Explicit style requests take priority. Do not treat the fictional reference characters as facts about the user, or infer age and gender from a name or photograph. For a personal character, use only relevant information the user has actually stated and clarify unresolved essentials.
| Decision | Original · flat mascot | Anime · soft-vinyl character |
|---|---|---|
| Main form | One continuous solid-color geometric body | Large rounded head with a tiny collar or body |
| Surface | Crisp flat fills, no shading or texture | Matte or restrained satin toy material and soft light |
| Reference | The chosen cell in one of the two shape atlases | The primary soft-vinyl reference and relevant character examples |
| Personal details | No age or gender needed for a shape mascot | Follow the explicit character brief; clarify missing personal essentials |
Prepare one complete brief without generating
For a fictional practice mascot, choose observed atlas cell A5: a blue rounded triangle with three recognizable sides and softened corners. Request two contrasting solid eyes, a plain background and clear edge space. Read SKILL.md, references/style-guide.md and inspect references/images/original-observed.png before selecting it as a generation reference. The extended atlas contains proposed design relatives, not an official catalog.
Record the requested count, silhouette, eye geometry, dominant color, background and destination. Write the generation prompt in English. The example below is a preparation brief, not a generated asset. Attach only the relevant atlas, explicitly assign its contour role and request no grid, cell labels or extra characters. Do not include unrelated personal history in the prompt.
Create one square profile avatar. Use observed atlas cell A5 as the contour reference: a rounded triangle with three readable sides and soft corners. One blue flat body, exactly two contrasting solid eyes, plain light background and clear edge space. No mouth, nose, pupils, shading, texture, text, grid or extra characters. Keep both eyes inside a circular crop.Check tool access before starting an image task
If image_gen is unavailable, stop at the reviewed brief and explain that generation has not occurred. Do not substitute a paid API or present a reference image as a new result. When the user authorizes generation, use the package’s reference-input rules and preserve each completed output before any follow-up. One individual avatar uses one call; a comparison sheet is one asset, not several independent avatar files.
A failed or ambiguous task needs investigation before retrying. For local references, inspect them first and pass only the permitted relevant files. Do not mix local-reference and recent-conversation reference mechanisms. This guide can be followed through brief preparation without invoking any generation service.
Review the delivered pixels at full size and at 48 pixels
- Compare the result with the selected mode. Original needs a recognizable uninterrupted shape and flat fills; Anime needs toy proportions, molded hair and a restrained soft-vinyl surface. A color change alone does not establish a new silhouette.
- Count exactly two solid eyes. Check for unwanted pupils, highlights, mouth, nose, text or extra characters. View the actual image at 48 pixels and inside a circular profile crop; both eyes must remain visible and distinct.
- Check the requested character, palette, count and dimensions. Preserve the approved image, exact non-sensitive brief, prompt, reference roles and actual output path. A checklist is a record of visual inspection, not an automatic quality certificate.
Keep a continuation separate from the finished delivery
An Original-to-Anime continuation preserves the approved mascot’s dominant color, eye arrangement and rounded motifs while adding a separately resolved character brief. It is another generation, not an included automatic conversion. Show the finished Original first and generate the continuation only when requested. Do not claim a locally prepared prompt proves the eventual image quality.
Prepare a source-photo portrait edit when preserving a real face matters