Back to prompts
Selfie transform
Default Image · 1 credit · 4:5

Selfie → watercolor portrait

Turn your selfie into a soft watercolor study.

Example output: a soft watercolor portrait of a young woman with freckles, wet-on-wet washes on textured paper. This is what your selfie looks like after running the prompt through MCP Media Engine img2img.
job: mjob_8cf55b8d-f7e2-4651-81ac-625380373e39 · 428,911B webp

The prompt

Verbatim. Paste this into your agent — Claude, ChatGPT, Cursor, or Hermes — and it will submit the job via MCP Media Engine.

transform into a loose watercolor portrait, soft wet-on-wet washes, paper texture, gentle skin tones, painterly, no text, no words, no logos

Run it in 3 steps

01 · install + key

Wire MCP Media Engine into your agent

Get a key at app.mcpmediaengine.com and set the env var.

npx -y mcp-media-engine

# set MEDIAENGINE_API_KEY in your env
# get a key at https://app.mcpmediaengine.com
02 · attach photo

Drop your image into the chat

Selfie, pet, product, or room — attach to your agent or paste any image URL.

# either:
# • attach the image to your agent chat
# • paste any image URL in the prompt

input_image_url: "https://your-image-url.jpg"
denoising_strength: 0.65  # recommended for img2img
03 · paste prompt

Run the prompt

The agent submits the job and returns the asset URL.

prompt: "transform into a loose watercolor portrait, soft wet-on-wet washes, paper texture, gentle skin tones, painterly, no text, no words, no logos"
aspect_ratio: "4:5"
model: "nano-banana"

Model + credits

credits per call

1 credit

aspect ratio

4:5

Live model catalog: /models. Credit costs surfaced at request time and echoed in the response.

Works in

This prompt is designed to run in any of these MCP-compatible agents. The wiring is the same — only the agent UI changes.

Claude
ChatGPT
Cursor
Hermes

Best for

IG reach posts — selfie transforms are the highest-engagement pillar.

Policy: use your own photos — don't upload photos of other people without their consent. No celebrities, no third-party faces. Example imagery on this page is generated via MED itself; job IDs are listed in the image caption.

Frequently asked

Where did the example image come from?
Generated via POST /api/jobs using a live MCP Media Engine key. Job ID mjob_8cf55b8d-f7e2-4651-81ac-625380373e39 (Default Image, 4:5). The example is text-to-image; running this with your own selfie uses the same prompt plus input_image_url + denoising_strength 0.65.

Related prompts

Learn more