Inspired by Zelrik#0969, who pointed out this fine workflow. Recursively using images can improve results.
Original prompt: beautiful woman with green hair and bright pink eyes
[URL of selected] , beautiful woman with green hair and bright pink eyes, gothic, night --ar 9:16
[URL of selected] , beautiful woman with green hair and bright pink eyes, gothic, night, rule of thirds --ar 6:4
While you can do this by right-clicking the image, getting the url, pasting it, and building your prompt string, you can simplify this process if you know how to do your own Discord bot (or have a friend who does).
Demonstration of Echo Llama Discord Bot. It listens for a MidJourney image response, then echos that image's URL and prompt back, for easy copy/paste to the Imagine Prompt.
Code to use in your Discord Bot:
For more advanced code, see https://github.com/Fledgewing/MidJourneyHelperBot/blob/main/index.js and https://wiki.artmechanicum.com/wiki/Help#What_is_MidJourneyHelperBot? by Temporalis#3311.
const Discord = require('discord.js'); const client = new Discord.Client({ intents: ["GUILDS", "GUILD_MESSAGES"] }); const settings = require('./settings.json'); client.on('ready',() => { console.log("online"); }); client.on('messageCreate', message => { if (message.author === client.user) return; if (message.author.username === "Midjourney Bot") { var str = message.content; var sub = str.match(/^\*\*.*\*\*/); if (sub) { sub = sub[0]; sub = sub.substr(2); sub = sub.substr(0, sub.length - 2) message.attachments.forEach(attachment => { const ImageLink = attachment.proxyURL; message.channel.send(`\`${ImageLink} , ${sub}\``); }); } } }); client.on('error', function (err) { console.log('Global error handler called:\n'); if(err) {console.log(err);} }); client.login(settings.token);