S Seedance How To 中文 Try Seedance ↗
← All articles
ToolsEnglish

Ultimate AI Image/Video Generator: End-User Guide

A practical workflow for generating images and videos with the Cyberbara API skill.

Ultimate AI Image/Video Generator: End-User Guide (English)

This guide is for regular users who want to generate images with the ultimate-ai-image-video-generator skill, without digging into internal implementation details.

It covers:

  • How to install the skill
  • How to get your CyberBara API key
  • How to write prompts that generate better results
  • Typical use cases
  • A complete first-image workflow

1. What This Skill Does

The skill wraps CyberBara Public API v1 into a practical CLI flow:

  1. Discover models
  2. (Optional) upload reference images
  3. Quote credit cost before submission
  4. Submit image/video generation tasks
  5. Poll task status until final output URL is ready

Core entrypoint:

python3 scripts/cyberbara_api.py --help

Service endpoint:

https://cyberbara.com

[Image Placeholder 1] Workflow diagram: models -> quote -> generate -> wait -> output

2. How to Install the Skill

2.1 If the skill is already available

If your environment already lists ultimate-ai-image-video-generator in AGENTS.md, you can skip installation.

2.2 Install via the Skill Installer (recommended)

Use the system skill-installer scripts:

# List installable skills
python3 ~/.codex/skills/.system/skill-installer/scripts/list-skills.py

# Install from openai/skills (example path format)
python3 ~/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py \
  --repo openai/skills \
  --path skills/.curated/ultimate-ai-image-video-generator

Or install from a direct GitHub URL:

python3 ~/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py \
  --url https://github.com/<owner>/<repo>/tree/<ref>/<path-to-skill>

After installation, restart Codex so the new skill is loaded.

[Image Placeholder 2] Screenshot: skill list + successful install output

3. How to Get a CyberBara API Key

  1. Open https://cyberbara.com/settings/apikeys
  2. Create a new API key
  3. Copy it once and store it safely
  4. Export it in terminal:
export CYBERBARA_API_KEY="<your_api_key>"

The CLI resolves API key in this order:

  1. --api-key
  2. CYBERBARA_API_KEY
  3. ~/.config/cyberbara/api_key
  4. interactive prompt

[Image Placeholder 3] Screenshot: CyberBara API key page

4. First Image in 5 Minutes (Step-by-Step)

Set your skill directory first:

SKILL_DIR="$HOME/.codex/skills/ultimate-ai-image-video-generator"

Step 1: Check available image models

python3 "$SKILL_DIR/scripts/cyberbara_api.py" models --media-type image

Step 2: Quote credits before generation

python3 "$SKILL_DIR/scripts/cyberbara_api.py" quote --json '{
  "model":"nano-banana-pro",
  "media_type":"image",
  "scene":"text-to-image",
  "prompt":"A cute kitten dancing, 3D cartoon style, dynamic pose, clean background",
  "options":{"aspect_ratio":"16:9","resolution":"1k"}
}'

Step 3: Generate the image task

python3 "$SKILL_DIR/scripts/cyberbara_api.py" generate-image --json '{
  "model":"nano-banana-pro",
  "scene":"text-to-image",
  "prompt":"A cute kitten dancing, 3D cartoon style, dynamic pose, cinematic lighting, cheerful mood",
  "options":{"aspect_ratio":"16:9","resolution":"1k"}
}'

By default, the CLI asks for explicit confirmation before spending credits.

Step 4: Wait for final result

python3 "$SKILL_DIR/scripts/cyberbara_api.py" wait --task-id <TASK_ID> --interval 5 --timeout 900

When status becomes success, use the output URL from output.images.

[Image Placeholder 4] Screenshot: pending -> success wait logs and final image URL

5. How to Prompt Better (Practical Template)

Use this prompt formula:

Subject + Action + Style + Camera/Framing + Lighting + Background + Quality cues

Example:

A cute orange kitten dancing on two legs, 3D cartoon style, full-body shot, soft studio lighting, clean pastel stage background, high detail, joyful expression

Prompt tips that improve consistency

  • Be explicit about style: 3D cartoon, anime, photorealistic, pixel art
  • Be explicit about framing: full body, close-up, wide shot
  • Be explicit about scene control: plain background, city street, indoor studio
  • Keep one main subject for stable output
  • Avoid conflicting instructions in one sentence

Resolution and aspect ratio examples

  • options.resolution: 1k, 2k, 4k
  • options.aspect_ratio: 1:1, 16:9, 9:16 (depending on model support)

[Image Placeholder 5] Side-by-side: weak prompt vs strong prompt result

6. Typical Use Cases

6.1 Social media creatives

  • Goal: fast hero images and thumbnails
  • Suggested setup: text-to-image, 16:9, 1k, vivid style prompts

6.2 Character concept exploration

  • Goal: generate many style variants of one character
  • Suggested setup: batch requests via --file

6.3 Product marketing visuals

  • Goal: e-commerce hero shots and campaign scenes
  • Suggested setup: image-to-image with reference product image in options.image_input

6.4 Storyboard keyframes for video pre-production

  • Goal: quickly draft scene keyframes before full video generation
  • Suggested setup: consistent scene template + multiple shot prompts

[Image Placeholder 6] Grid collage: 4 use-case outputs

7. Useful Commands Cheat Sheet

# Models
python3 "$SKILL_DIR/scripts/cyberbara_api.py" models --media-type image
python3 "$SKILL_DIR/scripts/cyberbara_api.py" models --media-type video

# Optional upload for image-to-image
python3 "$SKILL_DIR/scripts/cyberbara_api.py" upload-images ./ref1.png ./ref2.jpg

# Credit and billing
python3 "$SKILL_DIR/scripts/cyberbara_api.py" quote --json '{...}'
python3 "$SKILL_DIR/scripts/cyberbara_api.py" balance
python3 "$SKILL_DIR/scripts/cyberbara_api.py" usage --limit 20

# Generate and track
python3 "$SKILL_DIR/scripts/cyberbara_api.py" generate-image --json '{...}'
python3 "$SKILL_DIR/scripts/cyberbara_api.py" generate-video --json '{...}'
python3 "$SKILL_DIR/scripts/cyberbara_api.py" task --task-id <TASK_ID>
python3 "$SKILL_DIR/scripts/cyberbara_api.py" wait --task-id <TASK_ID> --interval 5 --timeout 900

8. Common Issues and Fixes

  • api_key_required / invalid_api_key
    • Verify key value and environment export
  • insufficient_credits
    • Run quote first, then recharge if needed
  • scene_not_supported
    • Check models output and use a supported scene
  • invalid_request
    • Ensure required fields exist and public options are inside options.*
  • task_not_found
    • Re-check task_id and environment/domain

9. Next Step

If you want, the next version of this guide can include:

  1. A full image pack for Placeholder 1-6
  2. A beginner prompt library (30+ ready-to-copy prompts by use case)