mcpField
T
transform-generate-image-with-transloadit

transform-generate-image-with-transloadit

github.com/transloaditCreative & media591 installsUpdated 1w ago
Repository

About transform-generate-image-with-transloadit

A skill for generating images with the Transloadit CLI, using prompt-only or image-guided inputs. It documents model selection, output options, and how to handle unavailable models.

Key features

  1. 01Prompt-based image generation
  2. 02Generation guided by input images
  3. 03Flare and Sunburst model selection
  4. 04Output path and dimensions options
  5. 05Assembly troubleshooting command

Use cases

  1. 01Create an image from a text prompt
  2. 02Guide a generated scene with reference images
  3. 03Choose precision-focused generation with Sunburst

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/transloadit/skills --skill transform-generate-image-with-transloadit

After installing, type/transform-generate-image-with-transloaditin Claude Code to confirm it loaded.

Details

From the project README

Run

Use the image generate intent with Images 2.5 Flare for quick image generation from a prompt. Pass the model explicitly so an older installed CLI does not select its previous default.

npx -y @transloadit/node image generate \
  --model openai/gpt-image-2.5-flare \
  --prompt 'A minimal product photo of a chameleon on white background' \
  --output ./out.png

Run With Images 2.5 Sunburst

Use Sunburst when the user asks for it or prioritizes precision over latency. Keep Flare as the default for everyday generation. Preserve any explicitly requested older model instead of silently substituting a newer one.

npx -y @transloadit/node image generate \
  --model openai/gpt-image-2.5-sunburst \
  --width 1024 \
  --height 1024 \
  --prompt 'A ceramic coffee mug on a white seamless studio background' \
  --output ./out.png

Run With Input Images

You can also guide generation with one or more input images. Prefer meaningful filenames and refer to them in the prompt.

npx -y @transloadit/node image generate \
  --model openai/gpt-image-2.5-flare \
  --input ./person1.jpg \
  --input ./person2.jpg \
  --input ./background.jpg \
  --prompt 'Place person1.jpg feeding person2.jpg in front of background.jpg' \
  --output ./out.png

Notes:

  • Use --model openai/gpt-image-2.5-flare by default, or --model openai/gpt-image-2.5-sunburst for precision-focused work.
  • Explicit --model openai/gpt-image-2 and Google Nano Banana selections remain supported. The older gpt-image-2 spelling still selects GPT Image 2, not Images 2.5.
  • Repeated --input values are bundled into a single /image/generate assembly.
  • Prompt-only generation still works without any --input.
  • Without --output, prompt-only and multi-input runs default to the current working directory.

Debug If It Fails

npx -y @transloadit/node assemblies get <assemblyIdOrUrl> -j

Notes:

  • Images 2.5 requires the matching API2 backend rollout. If a model is unavailable or account-gated, report that limitation and confirm an alternative with the user; do not silently change an explicit model choice or retry indefinitely.

Comments

Markdown supported