Krea AI — Image, Video & Enhancement Generation

Generate images, videos, upscale/enhance images, and train LoRA styles using the Krea.ai API. Supports 20+ image models (Flux, Imagen, GPT Image, Ideogram, Seedream...), 7 video models (Kling, Veo, Hailuo, Wan), and 3 upscalers (Topaz up to 22K).

IMPORTANT: Do NOT invent model names. Run

list_models.py

list_models.py --json

--model /generate/image/google/imagen-4-ultra

Usage

Scripts are in the

scripts/

uv run

Generate image:

uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "your description" --filename "output.png" [--model nano-banana-2] [--width 1024] [--height 1024] [--api-key KEY]

Generate video:

uv run ~/.codex/skills/krea/scripts/generate_video.py --prompt "your description" --filename "output.mp4" [--model veo-3.1-fast] [--duration 5] [--aspect-ratio 16:9] [--api-key KEY]

Enhance/upscale image:

uv run ~/.codex/skills/krea/scripts/enhance_image.py --image-url "https://..." --filename "upscaled.png" --width 4096 --height 4096 [--enhancer topaz-standard-enhance] [--api-key KEY]

Train a LoRA style:

uv run ~/.codex/skills/krea/scripts/train_style.py --name "my-style" --urls-file images.txt [--model flux_dev] [--type Style] [--trigger-word "mystyle"] [--api-key KEY]

List available models:

uv run ~/.codex/skills/krea/scripts/list_models.py [--type image|video|enhance]

Run a multi-step pipeline:

uv run ~/.codex/skills/krea/scripts/pipeline.py --pipeline pipeline.json [--api-key KEY]

Check job status:

uv run ~/.codex/skills/krea/scripts/get_job.py --job-id "uuid" [--api-key KEY]

Important: Always run from the user's current working directory so files are saved where the user is working.

Default Workflow (draft → iterate → final)

Goal: fast iteration without burning CU on expensive models until the prompt is right.

• Draft (cheap/fast): use

  --model z-image

  or

  --model flux-1-dev

  (3-5 CU, ~5s) for quick iteration

  uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "<draft prompt>" --filename "yyyy-mm-dd-hh-mm-ss-draft.png" --model flux-1-dev
  

• Iterate: adjust prompt, keep trying with cheap models

• Final (high quality): switch to

  --model gpt-image

  or

  --model nano-banana-pro

  uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "<final prompt>" --filename "yyyy-mm-dd-hh-mm-ss-final.png" --model nano-banana-pro
  

Available Models

Models, CU costs, and accepted body fields are fetched live from the Krea API's OpenAPI spec (

/openapi.json

list_models.py

uv run ~/.codex/skills/krea/scripts/list_models.py                     # all models with params
uv run ~/.codex/skills/krea/scripts/list_models.py --type image         # image models only
uv run ~/.codex/skills/krea/scripts/list_models.py --json               # machine-readable

Short aliases (e.g.

flux

flux-1-dev

Model selection guidance

Map user requests for images:

• "fast", "quick", "cheap" →

  flux-1-dev

  or

  z-image

• "high quality", "best" →

  nano-banana-pro

  or

  gpt-image

• "text in image", "typography" →

  ideogram-3

• "photorealistic" →

  seedream-4

  or

  nano-banana-pro

• No preference →

  nano-banana-2

Map user requests for video:

• "fast" →

  hailuo-2.3

• "cinematic", "high quality" →

  veo-3.1

• "with sound", "with audio" →

  veo-3

  with

  --generate-audio

• No preference →

  veo-3.1-fast

Enhancers:

topaz-standard-enhance

topaz-generative-enhance

topaz-bloom-enhance

Image Generation Parameters

--model

nano-banana-2

--prompt

--filename

--width

--height

--aspect-ratio

--resolution

--seed

--image-url

--style-id

--style-strength

--batch-size

--steps

--guidance-scale

--quality

--output-dir

--api-key

Video Generation Parameters

--model

veo-3.1-fast

--prompt

--filename

--duration

--aspect-ratio

--start-image

--end-image

--resolution

--mode

--generate-audio

--output-dir

--api-key

Enhancement Parameters

--enhancer

topaz-standard-enhance

--image-url

--filename

--width

--height

--enhancer-model

--creativity

--face-enhancement

--sharpen

--denoise

--scaling-factor

--output-format

--output-dir

--api-key

LoRA Training Parameters

--name

--model

flux_dev

--type

Style

--urls

--urls-file

--trigger-word

--learning-rate

--max-train-steps

--batch-size

--timeout

--skip-validation

--output-dir

--api-key

Training requires 3-2000 images. The script validates all URLs before submitting. Training takes 15-45 minutes. On completion, the style ID is printed to stdout and a

training-manifest.json

--output-dir

Use the style ID with

--style-id

generate_image.py

uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "mystyle product on white background" --style-id "style_abc123" --model flux-1-dev --filename "branded.png"

API Key

Scripts check for API key in this order:

1. --api-key

   argument (use if user provided key in chat)

2. KREA_API_TOKEN

   environment variable

If neither is available, the script exits with an error message.

Preflight + Common Failures

Preflight:

• command -v uv

  (must exist)

• test -n "$KREA_API_TOKEN"

  (or pass

  --api-key

  )

Common failures:

• Error: No API key

  → set

  KREA_API_TOKEN

  or pass

  --api-key

• 402 Insufficient credits

  → top up compute units at https://krea.ai/settings/billing

• 402 This model requires a higher plan

  → model needs a paid plan upgrade at https://krea.ai/settings/billing

• 429 Too many requests

  → concurrent job limit reached; scripts auto-retry up to 3 times with backoff

• Job failed

  → check prompt for content moderation issues, try different wording

Filename Generation

Generate filenames with the pattern:

yyyy-mm-dd-hh-mm-ss-name.ext

• Timestamp: current date/time in

  yyyy-mm-dd-hh-mm-ss

  (24h format)
• Name: descriptive lowercase text with hyphens (1-5 words)
• Extension:

  .png

  for images,

  .mp4

  for videos

Examples:

• Prompt "A cyberpunk cat" →

  2026-03-31-14-23-05-cyberpunk-cat.png

• Prompt "waves on a beach" →

  2026-03-31-15-30-12-beach-waves.mp4

Prompt Handling

For generation: Pass user's description as-is to

--prompt

For image-to-image: Use

--image-url

--prompt

For video from image: Use

--start-image

--prompt

Preserve user's creative intent in all cases.

Output

• Scripts download the result and save it to the current directory (or

  --output-dir

  )
• Script outputs the full path to the generated file
• Do not read the image/video back — just inform the user of the saved path
• If

  --batch-size

  > 1, files are saved as

  name-1.png

  ,

  name-2.png

  , etc.

Examples

Quick draft image:

uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "A serene Japanese garden with cherry blossoms" --filename "2026-03-31-14-23-05-japanese-garden.png"

High quality final:

uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "A serene Japanese garden with cherry blossoms, golden hour lighting" --filename "2026-03-31-14-25-30-japanese-garden-final.png" --model nano-banana-pro --resolution 4K

Image-to-image edit:

uv run ~/.codex/skills/krea/scripts/generate_image.py --prompt "transform to watercolor painting style" --filename "2026-03-31-14-30-00-watercolor.png" --image-url "https://example.com/photo.jpg" --model nano-banana-pro

Generate video:

uv run ~/.codex/skills/krea/scripts/generate_video.py --prompt "A majestic eagle soaring over snow-capped mountains at sunrise" --filename "2026-03-31-15-00-00-eagle-mountains.mp4" --model veo-3 --duration 8 --generate-audio

Upscale image to 4K:

uv run ~/.codex/skills/krea/scripts/enhance_image.py --image-url "https://example.com/photo.jpg" --filename "2026-03-31-15-10-00-upscaled.png" --width 4096 --height 4096 --enhancer topaz

Train a LoRA style:

uv run ~/.codex/skills/krea/scripts/train_style.py --name "acme-brand" --model flux_dev --type Style --trigger-word "acmestyle" --urls-file brand-images.txt --output-dir output/acme-brand

List models:

uv run ~/.codex/skills/krea/scripts/list_models.py --type image

Pipelines (Multi-Step Workflows)

For multi-step workflows (generate → enhance → animate, fan_out branching, template variables, parallel execution, resume, dry-run), see PIPELINES.md.

Quick example:

uv run ~/.codex/skills/krea/scripts/pipeline.py --pipeline '{"steps":[{"action":"generate_image","prompt":"a cat astronaut","filename":"cat"},{"action":"enhance","use_previous":true,"enhancer":"topaz-standard-enhance","width":4096,"height":4096,"filename":"cat-4k"}]}'