Destesi ImageGen

Generate images from a prompt with Stable Diffusion XL. Streamed previews, every image saved to your Drive library, scriptable API.

Open generatorRead the docs

Watercolor

“A red panda curled on a mossy branch, watercolor, soft pastel palette”

Cinematic

“Neon-lit alley in the rain, cinematic lighting, ultra-wide”

Studio

“Minimal product shot of a ceramic mug, studio softbox, cream backdrop”

Isometric

“Isometric floating island with waterfalls, vibrant colors”

One prompt in, streamed previews out — open the generator and try one.

How it works

1

Prompt it

Chat with the studio agent — “darker”, “make it vertical”, “another variation” — or use the classic prompt form.

2

Watch it stream

Live progress while the model renders — no spinner-and-pray. Iterate as soon as the preview lands.

3

It's in your Drive

Every image is registered in your Drive library automatically — one place for everything your workspace makes.

Two engines, one prompt box

Switch per generation — same gallery, same Drive library either way.

Fast

default

Stable Diffusion XL. Quick drafts, style exploration, and batch iteration. No setup — it just works.

High quality

via Connect

GPT image generation with your own OpenAI key, connected once in Destesi Connect. Sharper detail, better text, tighter prompt following.

Scriptable, end to end

Everything the dashboard does is one HTTP call away. Mint a personal access token in your account, and generate from CI, a cron job, or your own app.

  • Synchronous — one POST, image URL back.
  • Streaming — SSE progress events for live UIs.
  • Tunable — size, steps, guidance, negative prompts.
Full API reference →
terminal
curl https://api.imagegen.destesi.io/v1/generate \
  -H 'Authorization: Bearer idn_pat_...' \
  -H 'X-Destesi-Workspace: my-team' \
  -d '{
    "prompt": "a watercolor fox in a misty forest",
    "negative_prompt": "blurry, text, watermark",
    "width": 1024, "height": 1024
  }'

What's inside

  • Chat studio. Describe an image, then iterate in plain language.
  • Prompt studio. Classic form with live streamed previews.
  • Gallery. Every render, browsable across your workspace.
  • Two engines. Fast SDXL, or high-quality GPT via Connect.
  • Drive library. Every image lands in your Drive automatically.
  • API tokens. Personal access tokens for scripted generation.

Type a prompt. Keep the good ones.

Free to try in your workspace — everything you make is saved to Drive.

Open generatorRead the docs