DALL·E · images
DALL·E integration in your own tool
DALL·E integration means a user in your system writes or picks a brief and gets an image back, which you store on the case. It makes sense for mockups, simple illustrations and placeholder assets. It does not make sense as an “AI section” on an internal warehouse system. Store prompt, model and version on the asset. Agree what the customer may use the image for. Cap spend per user. A human approves before it goes into a customer-facing document. DALL·E if you are already on the OpenAI stack and want a simple API. [Stable Diffusion](/articles/integrations/stable-diffusion-integration) if you want to host yourself or control more.
DALL·E is an OpenAI socket next to GPT. Keep it in the same engine layer, the same cap and the same log. AI when it is a feature.
Rules
- Store prompt + model + seed/version on the asset
- Rights: what may the customer use the image for
- Spend per user — or it becomes a toy
- A human approves before it goes into a customer-facing document
DALL·E
- You are already on the OpenAI stack
- You want a simple API
- You want to move this sprint
Stable Diffusion
- You want to control model, LoRA and style
- You want to control whether data leaves
- You have volume or a self-host requirement
Prompt
Menneske godkender før udsendelse
What to do now
Describe one place the image is part of the delivery — mockup, illustration, placeholder. Not “we also want AI images”. Tell us the job.
FAQ about DALL·E integration
DALL·E or Stable Diffusion?
DALL·E if you are already on the OpenAI stack and want a simple API. Stable Diffusion if you want to host yourself or control more.
May the customer use the image commercially?
That is decided by OpenAI’s terms and your agreement with the customer. Write it on the asset, not in a Slack thread.
Must every image be approved?
Yes, before it is customer-facing. Internal placeholders can have a softer track — write that in the brief.