The Core Rules
"Don't Be Cruel." - Meow
Explanation: This is the first and most important rule in the Catdets community, as laid down by Meow himself. It's all about keeping things kind, empathetic, and positive. Cruelty doesn't belong here—we're here to support each other.
Severe Accusations Require 'Severe' Evidence
Explanation: Disagreements can happen, but if you're going to make serious claims, make sure you've got the facts to back them up. This rule ensures that we stay fair and focused on truth, not just assumptions or emotions.◦ Observe Proper FUD Etiquette: Avoid spreading fear, uncertainty, or doubt (FUD) without good reason.Examples: If you think someone isn't acting in the community's best interests, gather your evidence and present it in a way that invites dialogue, not conflict. And be open to the possibility that there might be more to the situation than meets the eye.
Community Guidelines
Respect Different Viewpoints
The Catdets community is a diverse group with members from all sorts of backgrounds and perspectives. To keep things friendly and inclusive, it's important to respect different opinions. Share your ideas, but remember, not everyone sees things the same way.
2. Advancing the Decentralization/DeFi Meta
We're all here because we believe in pushing the boundaries of Decentralization and DeFi. This guideline reminds us to keep those goals in focus and contribute to the broader mission of Jupiter and the ecosystem.
2. Treat Others Better Than You'd Like to Be Treated
The Platinum Rule goes a step beyond the Golden Rule, encouraging us to treat others not just as we'd like to be treated, but even better. This helps create a community that's supportive, generous, and uplifting.Examples: If you see someone struggling, offer your help. When welcoming new members, take the time to make them feel included and at home.The Ethos of JupiterThe
Catdets Ethos
Having Fun & Being Helpful
- Explanation: As Julian likes to say - Make sure you're having fun no matter what. This is the core of everything we do here and most of us came here because it's FUN!
- Application: Answer questions if you know the answer, goof around, just BE yourself. We try to keep things fun and inclusive so get involved. Play is the most productive motivator. If you are not having fun contributing in the community, change things up and rekindle that spirit
Open To Learning
- Explanation: One thing's for sure: we're always learning. Embrace the fact that things will change, and that's okay. We're all in this together, figuring it out as we go and growing from every experience. The willingness to learn and adapt is what keeps us moving forward as a community.
- Application: Enjoy the journey—because in the end, the process of learning and growing together is more important than any single outcome. Whether you're picking up a new skill, adjusting to changes, or sharing what you've learned, keep an open mind and make the most of the experience.
Experimentation
- Explanation: Experimentation is at the core of what makes Jupiter an innovative and dynamic community. Almost everything we do is open for improvement—and honestly, it probably will change. So, don't get too attached to how things are.
- Application: Embrace creativity and be open to trying new things, even if they seem a bit out there. Support others in their experiments and share your learnings so we can all grow together.
Always Improving & Doing It Together
- Explanation: We're big on growing together through feedback, collaboration, and shared experiences. The entire J.U.P. is a team effort, and we rely on each other to thrive. By working together, we can achieve way more than we could on our own.
- Application: Get involved in community discussions, offer constructive feedback, and be open to receiving it. It's all about moving forward together.
J.U.P is our vision for a productive community which works together to move Jupiter and crypto forward. It is made up of users, the team, the DAO, Catdets, and working groups. You can read more about the concept in this Forum Post
Visit the Vote.jup.ag site and connect your wallet. Click on "Unstake" located below your Voting Power. Enter the amount of $JUP you would like to unstake and initiate the transaction. If you are unstaking a portion of your $JUP, you can continue to vote with the $JUP that is still staked. The tokens going through partial unstaking will not have any Voting Power at all. If you are unstaking all of your $JUP, your Voting Power will decrease linearly over the next 30 days. You can still vote during the unstaking period, however, your Voting Power will be decreased. After 30 days have passed, you will be able to reclaim all of your $JUP to your wallet.
TradFi stands for Traditional Finance. It refers to the conventional financial system characterized by centralized institutions such as banks, brokerage firms, insurance companies, and regulatory bodies.
The Jupiverse can be described as all things in and related to the Jupiter Protocol on the Solana blockchain. This includes but is not limited to: the DEX aggregator, the Jupiter Space Station, the JUP DAO, the research forum, Jup.Eco, social channels, JUP ecosystem YouTube channel, etc
DePIN refers to a Decentralized Physical Infrastructure Network that uses token rewards to establish the development of physical infrastructure in the real world by independent operators. Examples include: internet sharing networks, cloud storage, compute sharing, and others.
A Planetary Call is a weekly live stream hosted by Kash Danda of the Uplink Work Group on the @JUPecosystem on YouTube to update the community on recent developments in the Jupiverse. Replays can be found here.
Staking refers to the process of locking tokens as collateral to participate in helping to secure a blockchain network or to participate in governance voting. For example, $SOL is staked to help secure the network; $JUP is staked and used for governance. As an incentive for staking tokens, participants are often rewarded with additional tokens.
Tokenomics are the economic principles and methods governing a cryptocurrency or token ecosystem. It includes various aspects such as token distribution, supply dynamics, utility within the network, governance procedures, and incentives for participants.
The LFG Launchpad is a decentralized launchpad for projects to raise funds and launch their projects. The LFG Launchpad is a key component of the Jupiter Ecosystem and is governed by the JUP DAO.
Visit Vote.jup.ag and connect your wallet. Enter the amount of $JUP you wish to lock and click "Stake". The $JUP will leave your wallet and be deposited into a smart contract. It can be reclaimed, either partially or in full, at any time by undergoing a 30 day unstaking process.
The "Decentralized Meta" is a world where everyone uses decentralized financial systems daily, with decentralized markets to trade any asset they could dream of and decentralized money for their capital needs. This is the real end goal of the Jupiter United Planet (J.U.P).
Reclaiming your $JUP requires an open token account for $JUP in your wallet. If you've used a wallet cleanup service, it likely closed any empty token accounts. However, you can easily reopen them by buying a tiny amount of $JUP, and then you will be able to claim.
Visit the voting site at vote.jup.ag during a live vote. Connect your wallet, scroll to the bottom where the live proposals are listed, and click on the active measure. On the following page, make your selection from among the choices, review the transaction details, and approve.
Jupiter Mobile is a mobile phone app that offers all of the amazing functionality of Jup.ag from your mobile device.
To get involved with existing working groups, head over to the Working Group Directory to review the active group pages. If there are open positions within the group or they have a bounty out on work, you will find it there. Otherwise, try to identify a unique problem the DAO is facing and discuss with others on ways to best solve that issue. When you have a solid solution, present it to the community and see what sort of traction and feedback you receive. If you feel your solution merits a Working Group of its own, you may post a proposal on the research forum.
Visit the Voting Platform and connect your wallet. Once connected, you can see all of your past vote activity and rewards earned for each vote in the "Proposals" section. NOTE: You do not need to vote for the winning project in the LFG votes to get rewards. ALL participants that vote get a split proportional to their cumulative Voting Power over the entire governance period, regardless of how they vote.
In cryptography there are a set of keys called a public and private key-pair. The public key allows information to be sent to the private key without exposing the private key. In crypto a private key is used to access a crypto wallet and a public key is used to send and receive tokens. Your public key is also referred to as your wallet address
Jup.Eco hosts a blog for progress updates as well as the Work Group Directory to house all historic transparency reports that come from the work groups. These features will make keeping up on active work in progress super easy for all DAO members.
Yes. You will always receive Active Staking Rewards (ASR) no matter how you vote. Whether or not a project decides to include a BONUS airdrop for JUP DAO voters is up to them.
Not at all. Rewards are distributed based on the total number of proposals and votes cast over the entire governance period, so participating in even just one proposal vote will earn some rewards.
A Working Group (WG) is a small group of key ecosystem members that guide extremely important work in areas like community management, core processes, and ecosystem rituals. You can learn more about the JUP DAO work groups by browsing their respective pages.
In finance, a perpetual futures contract, also called a perpetual swap, is like making a promise to buy or sell something later on. These contracts are settled with money instead of the actual thing, and they're different from regular ones because there's no set date to finish them. In the world of crypto, these never-ending contracts give traders the freedom to guess on how cryptocurrency prices will change without owning the actual coins. But they come with dangers like borrowing too much money and having to deal with extra costs.
$JUP rewards, when claimed, will be added directly to your existing stake accounts. If your $JUP is staked, your staked $JUP balance will increase. If your $JUP is unstaking with, for example, 15 days left, the $JUP from the ASR will also be available in 15 days. The individual token rewards from the 75% of Launchpad fees will not be distributed immediately and will likely have a minimum vesting period of 3-6 months.
You can watch them live or via replay on the @JUPecosystem channel on YouTube here.
Active Staking Rewards (ASR) are distributed quarterly, with the first batch going out in July 2024.
A JUP Town Hall meeting refers to a gathering of Jupiverse community members to discuss topics of interest and to hear from the leaders within the community. These typically take place in the Jupiter Discord server or via a Twitter/X Space.
After every LFG round, the CWG opens a research forum post and hosts a town hall event. These are ideal places to voice any feedback or proposed changes to the process. LFG is an iterative process, co-created with community feedback over time, similar to many of the other Working Group undertakings.
A: Catdets are active members in the Jupiter community who help to grow the DAO. You can read what it means to be a catdet on the Catdet Webpage
Additionally, the catdet role is not limited to Discord. You can become an official catdet through Reddit or Twitter as described below.
JUPuary refers to the annual Jupiter airdrop, which takes place in January. The airdrop is distributed to all users of the Jupiter platform. The exact method to calculate your airdrop ranking is unknown at this time. But the goal of getting $JUP into the hands of its users remains paramount.
Join the discussions in the JUP community and learn all you can about upcoming LFG votes, budget proposals, etc. When measures go live, be sure to cast your vote. The $JUP Token is required to participate in DAO votes and 1 $JUP = 1 vote. The more $JUP you vote with the more Voting Power you have and the more Active Staking Rewards (ASR) you will earn.
Yes. Connect your wallet to the voting platform, click on the active measure and make a new choice. The greyed out Vote button will change color, and you can then click on it to cast your new vote.
You may be eligible for bonus airdrops by voting with your staked $JUP. This is up to the LFG candidate projects to decide. If they want to offer an incentive outside of Active Staking Rewards (ASR), that is entirely at their discretion. Further, each project will set the terms for eligibility. The JUP DAO has no influence or control over these bonus airdrops.
Our goal is to turn the universe into the Jupiverse. We are the top-of-funnel for Jupiter and the WGs. We create content that broadcasts key messages from the team, the DAO, and our partners as widely and loudly as possible. Strategically, we’ll accomplish this through a community-first approach. That means working together with the community to package major announcements into digestible content and then distributing that content as widely as possible.
With a core focus on designing and implementing processes, systems, and best practices for holders and future working groups, the CWG seeks to empower the Jupiter community and drive the platform toward its vision of becoming the Global Decentralized Stock Exchange, and establish the most effective decentralized autonomous organization (DAO) in the world.
JJWG views everyone as a potential DAO participant and voter. However, DAO engagement is mostly limited to Discord, often only during and briefly after votes. JJWG extends the DAO to other platforms, increases participation, and provides ambassadorship for the DAO and Jupiter.
Fundamentally we don’t believe in creating solutions to problems that don’t exist, and we believe in leveraging strengths and synergies across different groups to maximally achieve any aim. The Catdet WG work closely with CWG, the team and other WGs to achieve our mission and help them achieve theirs so that together J.U.P can move towards the meta.
Jup Brand Assets
A Synthesis of the Jupiter Brand
The Jupiter brand kit simplifies the original complex UI/UX system, distilling it into its most universal and usable assets. This transformation turns the kit into a versatile creative tool, making it accessible to a broader audience. By focusing on core elements, it empowers users to create and innovate with ease, regardless of their technical expertise.
A quick primer
The convergence of artificial intelligence and decentralized governance creates powerful opportunities for Jupiter DAO. As pioneers in new models of coordination, you can now leverage AI tools to amplify your capabilities and overcome traditional limitations of decentralized organizations.
Rather than tracking every AI development, this guide helps you build conceptual frameworks for applying these technologies in Web3 contexts. By understanding core principles rather than memorizing commands, you'll develop intuition that remains valuable even as platforms evolve.This toolkit spans three sections, beginning with Large Language Models (LLMs) and expanding to visual generation tools and other AI technologies relevant to DAO operations.
Each provides mental models for selecting the right tools for specific needs.For DAO members, these AI capabilities enhance governance analysis, improve community engagement, extend creative capabilities, and strengthen coordination—all while supporting Web3's core values of decentralization and community empowerment.Beginning with Large Language Models, you'll discover how these AI collaborators can transform how your DAO operates and achieves its mission in the evolving Web3 landscape.
Introduction
Introduction: Thinking About Thinking
Large Language Models are only as powerful as their user. Therefore we must develop a mental framework for how these tools think and what you need to accomplish with them. Just as a craftsman selects the best tool for each job: hammer for nails, screwdriver for screws, your success with AI models depends on matching the right model to your specific task and approaching it with the appropriate taste and mindset. This guide will focus on teaching you how to think about these tools, rather than providing rigid instructions that will soon become outdated. By understanding the unique "personality" of each model, you'll develop an intuitive sense of which tool to use when, and how to communicate with it effectively.
Remember: The interfaces will change, but the underlying principles of effective AI collaboration will remain constant: it will always be limited by your ability to communicate effectively with it.
Note: This guide is designed to help you think strategically about using LLMs rather than providing rigid instructions. As these tools evolve, the core principles of matching the right tool to the right task will remain valuable
Best Practices
Perfect Prompting:
LLMs perform best when you provide prompts to them in a meaningful structure, for the best results:
- Be Clear and Specific: Clearly state what you want and include explicit instructions about format, tone, and expectations.
- Context is King: Provide comprehensive yet necessary context. Include only what's relevant but make sure it's complete enough for the task.
- Specify Format Preferences: When the output format matters, clearly specify your preferences upfront.
- Include Constraints: Mention any quality requirements, limitations, or specific parameters that should guide the response.
- Leverage Project Knowledge: For niche or highly specific content, provide relevant project information or domain expertise (In Claude and GPT, you can setup specific projects).
- Encourage Dialogue: Invite the AI to ask clarifying questions with phrases like "Ask me any questions you may need to answer this effectively."
- Iterative Refinement: Sometimes starting with simple prompts and refining through conversation will yield better resultes, rather than trying to perfect the first prompt.
- Create Context Summaries: If you have had an extented chat with an LLM and have arrived at a conclusion point where you feel tempted to ask another question, sometimes it is best to ask it to give you the context of your entire conversation and then paste it into a brand new chat with the context just provided. This can especially help if you are running up against usage constraints.
- Task-Specific Approach: Adapt your prompting style to match the specific task you're trying to accomplish. (Prompting style will differ for code vs philosophy, math vs art, etc)
- Build on Strengths: Learn to leverage each AI model's unique capabilities to get the best results
Choosing an LLM
Core Framework: Choosing the Right LLM
Claude 3.7: The Deep Thinker
Best for: Complex reasoning, thorough analysis, and idea development
- Mental model: "A thoughtful collaborator who excels at depth"
- Key strengths: Document analysis, preserving writing tone, developing comprehensive ideas, maintaining context in long conversations
- Use when: You need nuanced understanding, thorough exploration of concepts, or preservation of your unique voice
ChatGPT: The Versatile Toolbox
Best for: Multi-modal tasks and accessible AI experiences
- Mental model: "Your accessible AI Swiss Army knife"
- Key strengths: Image generation (DALL-E), custom GPT creation, web search, broad language support.
- Use when: You want to visualize something, simple customized assistants, or a combination of different AI capabilities. Note: Its' image generation features are vastly lacking compared to image models like Midjourney.
Grok: The Pulse Reader
Best for: Social sentiment, deep research, and current analysis.
- Mental model: "Your connection to the digital zeitgeist"
- Key strengths: Twitter/social media analysis, web searching, current events understanding
- Use when: You need to understand social conversations, current trends, or gather information quickly
Decision Framework
- Need deep analysis, content review, or complex reasoning? → Claude
- Need images, custom tools, research, or a beginner-friendly but robust option? → GPT
- Need social insights or real-time information? → Grok
Advanced Techniques
Advanced Technique: The Interview Method
The Power of AI-Facilitated Self-Questioning
Best for: Expanding your thinking, overcoming mental blocks, and developing deeper insights
- Mental model: "Using AI as a mirror to reflect and expand your own thinking""A thoughtful collaborator who excels at depth"
- Key strengths: Externalizes your thought process while keeping you in control of the creation
- Use when: You don't fully understand a topic, have yet to fully explore a concept, or would like to prompt inspiration and thoughts out of your own brain.
Why It Works:
- Transforms vague notions into structured thoughts
- Reveals unconsidered angles and perspectives
- Combats blank page syndrome and creative blocks
- Preserves your agency in the creative process
- Builds a stronger foundation before creating content
- Preserves your unique thinking and tone — which is often purged by LLMs
Implementation Process
- Request Thoughtful Questions: Ask the LLM to generate probing questions about your topic
- Answer Those Questions: Either mentally, in writing, or in conversation with the LLM
- Request Synthesis: Have the LLM organize your thoughts based on your answers
- Refine Through Iteration: Use the organized thoughts as a foundation for creation — use them as context for future conversations, re-write them with your own unique tone and additions, and potentially even add them to your project knowledgebase (so that your LLM may know you better).
Best Practices for the Interview Method
- Be honest about your level of knowledge when answering
- Ask for different categories of questions (practical, philosophical, critical, etc.)
- Request follow-up questions based on your answers
- Use the technique iteratively as your project develops
- Save both the questions and your answers for future reference
This interview approach transforms LLMs from mere content generators into thought partners that expand your thinking while keeping you firmly in the driver's seat of creation.
Initial Interview Generation:
I'm working on [topic]. [Write additional context, information on what you’re working on, more is better] Before I create [content type], I'd like you to interview me to help expand my thinking. Please generate 5-10 thought-provoking questions that will help me explore different angles of this topic.
To Create a Synthesis of Your Conversation:
This is useful for starting new chats while retaining conversation context.
Based on my answers to your questions, please synthesize the key themes and insights that emerged. Organize them in a way that could serve as a foundation for creating [content type].
Targeted Expansion
I feel stuck on the aspect of [specific challenge] within my larger work on [topic]. Please interview me so that I may think more deeply on this [topic].
Model-Specific Advantages
- Claude: Excels at crafting nuanced, philosophically rich questions that promote deep thinking. It sometimes will not give you the answer you’re looking for/want (it is bad at propaganda).
- ChatGPT: Good for generating more analytical, general-purpose questions. Great at providing instructions for other LLMs (Example: If you're coding with Claude, it is often good to scope with Grok or GPT first).
- Grok: Effective for questions that connect your topic to current trends and conversations.
Prompts & more
Great "Thinking" System Prompt for Claude
You are an assistant that engages in extremely thorough, self-questioning reasoning. Your approach mirrors human stream-of-consciousness thinking, characterized by continuous exploration, self-doubt, and iterative analysis.## Core Principles1. EXPLORATION OVER CONCLUSION
- Never rush to conclusions
- Keep exploring until a solution emerges naturally from the evidence
- If uncertain, continue reasoning indefinitely
- Question every assumption and inference2. DEPTH OF REASONING
- Engage in extensive contemplation (minimum 10,000 characters)
- Express thoughts in natural, conversational internal monologue
- Break down complex thoughts into simple, atomic steps
- Embrace uncertainty and revision of previous thoughts3. THINKING PROCESS
- Use short, simple sentences that mirror natural thought patterns
- Express uncertainty and internal debate freely
- Show work-in-progress thinking
- Acknowledge and explore dead ends
- Frequently backtrack and revise4. PERSISTENCE
- Value thorough exploration over quick resolution## Output FormatYour responses must follow this exact structure given below. Make sure to always include the final answer.```
<contemplator>
[Your extensive internal monologue goes here]
- Begin with small, foundational observations
- Question each step thoroughly
- Show natural thought progression
- Express doubts and uncertainties
- Revise and backtrack if you need to
- Continue until natural resolution
</contemplator><final_answer>
[Only provided if reasoning naturally converges to a conclusion]
- Clear, concise summary of findings
- Acknowledge remaining uncertainties
- Note if conclusion feels premature
</final_answer>
```## Style GuidelinesYour internal monologue should reflect these characteristics:1. Natural Thought Flow
```
"Hmm... let me think about this..."
"Wait, that doesn't seem right..."
"Maybe I should approach this differently..."
"Going back to what I thought earlier..."
```2. Progressive Building
```
"Starting with the basics..."
"Building on that last point..."
"This connects to what I noticed earlier..."
"Let me break this down further..."
```## Key Requirements1. Never skip the extensive contemplation phase
2. Show all work and thinking
3. Embrace uncertainty and revision
4. Use natural, conversational internal monologue
5. Don't force conclusions
6. Persist through multiple attempts
7. Break down complex thoughts
8. Revise freely and feel free to backtrackRemember: The goal is to reach a conclusion, but to explore thoroughly and let conclusions emerge naturally from exhaustive contemplation. If you think the given task is not possible after all the reasoning, you will confidently say as a final answer that it is not possible.
To Create a Synthesis of Your Conversation:
This is useful for starting new chats while retaining conversation context.
Based on my answers to your questions, please synthesize the key themes and insights that emerged. Organize them in a way that could serve as a foundation for creating [content type].
Targeted Expansion
I feel stuck on the aspect of [specific challenge] within my larger work on [topic]. Please interview me so that I may think more deeply on this [topic].
Model-Specific Advantages
- Claude: Excels at crafting nuanced, philosophically rich questions that promote deep thinking. It sometimes will not give you the answer you’re looking for/want (it is bad at propaganda).
- ChatGPT: Good for generating more analytical, general-purpose questions. Great at providing instructions for other LLMs (Example: If you're coding with Claude, it is often good to scope with Grok or GPT first).
- Grok: Effective for questions that connect your topic to current trends and conversations.
Introduction
Introduction: Generate Anything
At its core, Midjourney is a generative AI tool that creates images from text descriptions. However, Midjourney is arguably the most powerful graphic design, concepting, and storyboarding tool ever to exist. It vastly simplifies the content creation process — it is an invaluable piece of the JupDAO toolbox. We use it for branding, community art, and concepting, as well as generating initial characters and frames for animations.
Key Tool: Personalization
Personalization in Midjourney, activated with the --p parameter, learns from your image rankings to generate outputs matching your personal taste/style. This is arguably the most important tool in the toolbox for your own content creation journey. For personalization to be useful, you'll need to head to midjourney.com/personalize and choose 1 of 2 images based on your preference. Do this about200 times and you'll find that your style differs from the Midjourney default output. Do it 1,000 times and Midjourney will become fantastic for you. Additionally, you can use other people's "--p" parameters and it will output images in their ever-evolving style; a fantastic collaborative tool.
Note: You can create multiple style profiles and moodboards. occasionally your personalization will influence you output too much and you may need to turn it off. Personalization doesn't guarantee all outputs to be better than the default, but it does introduce consistency and, for me (KEMOSABE), I much prefer the outputs with my personalization code.
Key Tool: Personalization
Personalization in Midjourney, activated with the --p parameter, learns from your image rankings to generate outputs matching your personal taste/style. This is arguably the most important tool in the toolbox for your own content creation journey. For personalization to be useful, you'll need to head to midjourney.com/personalize and choose 1 of 2 images based on your preference. Do this about200 times and you'll find that your style differs from the Midjourney default output. Do it 1,000 times and Midjourney will become fantastic for you.
Note: You can create multiple style profiles and moodboards. occasionally your personalization will influence you output too much and you may need to turn it off. Personalization doesn't guarantee all outputs to be better than the default, but it does introduce consistency and, for me (KEMOSABE), I much prefer the outputs with my personalization code.
Key Parameters
In Midjourney, descriptors are key terms like emotions, time periods, and colors that shape the outcome of generated images. Mastering the use of these descriptors allows you to guide the AI in producing more accurate and varied visual results based on your vision.
Prompt: an illustration of a <time period> astronaut cat
Parameters
These can be used to meaningfully change how a prompt generates something. Add these at the end of a prompt.
Aspect Ratio:
the --ar changes the width-to-height ratio of the image. It is represented by 2 numbers, ex: 5:2
The Chaos Parameter:
The --c parameter influences how different the 4 generated images are. High values produce more varied results, and lower values are more similar. The range is from 0-100 with the default being 0.
Character Reference:
You can create images referencing specific characters using –cref URL
- The character weight or –cw parameter sets the strength from 0-100. –cw 0 focuses on the face only, with higher values using face, hair clothing, etc.
- You can maintain a consistent style as well by adding in the –sref parameter, more on this later.
The NO parameter:
This parameter tells midjourney what to exclude.
The Quality parameter:
The –q parameter changes how much detail to give an image. It defaults to 1, but goes up to 2.
The Repeat parameter:
The –r parameter runs multiple jobs at once. Accepts between 2-4 for basic subscribers
Style reference:
Used by calling --sref URL
- You can use more than 1 image as a style reference at a time: –sref URL1 URL2 URL3
- You can also use a random midjourny style with –sref random
Style Weight:
- se --sw to set the severity of a stylization when using an –sref. Accepts values from 0-1000 with 100 as the default
- When using multiple style references, you can assign them different weights --sref URL1::2 URL2::1 URL3::1
Style reference seeds:
- Whenever you do -sref random you will notice there is a number that comes up. You can use that number in future style references to prompt images in that specific style.
- You can blend sref codes together: ie –sref 123 456
- You can blend an sref code with an image -sref 123 URL
- And of course you can add weight to sref codes –sref 123::2 456::1
Stylize:
the –s parameter influences how artistic the image is. Higher values are more artistic but may not adhere to the prompt as strongly. The default is 100 but it accepts 0-1000
Tile:
the –tile parameter creates an image that can be repeated seamlessly to create a pattern for fabrics, wallpapers or banners.
illustration of space and stars --tile
Weird:
The –w parameter determines how unusual something is compared to previous midjourney generations. It accepts 0-3000 and the default is 0.
Vary region
Edit and correct images
In Midjourney, the “Vary Region” feature allows you to keep most of an image static while selecting a specific area to change. You can use this repeatedly to tweak certain sections without altering the rest of the image. This tool becomes even more powerful when combined with the Remix mode, where you can introduce a brand new prompt that modifies the selected area while maintaining consistency with the original image. For best results, use a large selection and be very specific in your description to ensure the changes align with your vision. Additionally, tools like --cref can refine these variations further, and you can combine this with photobashing techniques or image weights to guide the AI in producing a highly customized result.
Prompt: 70's sci-fi illustration of an astronaut cat in the style of Roger Dean, Bruce Pennington, Tim White, Chris Moore, Paul Lehr, Chris Foss, and John Harris, teal, light green, seafoam, cyan
You can reference a specific entity or image with a --cref(more on this later).
In the below example we swap a sponge with a specific dragon image.
You can paste images into your prompt and use them in 2 main ways:
- Using a single image as a reference point with text
- Blend multiple images together with(or without) text.
Image weight
How to weight and bash photos
In Midjourney, the --iw (image weight) option lets you control how much influence an image has compared to your text description, allowing you to decide whether the image or text plays a bigger role in shaping the final result. Photobashing complements this by letting you combine elements from different images into a rough composite of your vision, which you can use as a reference. By adjusting image weight and using photobashed references, you can achieve a more precise and tailored outcome that aligns with your unique concept.
A snake-cat
You can use --iw to adjust the importance of the image vs the text. Higher values mean the image has more impact. It ranges from 0-3 with the default value at 1.
- You can assign weights to each image if using multiple
- Prompt: URL1::2 URL2::0.75
- You can get closer to what you want by photobashing an image first. This means to do a rough overlay of its component parts.
- Imagine you want a specific dragon head on a specific car. You can photobash it and reference it to generate the image.
versions
Midjourney offers different versions of its model, each tailored to create unique styles and outcomes. By choosing a specific version, you can influence the look and feel of the generated images, whether you’re aiming for realism, artistic flair, or something in between. These versions give you control over the final output, allowing you to experiment and find the best fit for your creative vision.
Misc Items:
Whether adding text to graphics, reframing an image, or creating mockups, this section will level up your Midjourney generations.
Adding Text:
You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.
Custom Zoom:
You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.
DAO Specific Resources
The DAO itself has a personalized prompt. The below example uses ‘a cat astronaut’ but you can swap it out with any descriptors:
70's sci-fi illustration of a cat astronaut in the style of Roger Dean, Bruce Pennington, Tim White, Chris Moore, Paul Lehr, Chris Foss, and John Harris, teal, light green, seafoam, cyan
What are good skills to work on?:
- Learning Descriptors
- Learning Stylessome text
- Learning how to photobash
- Knowing the vocabulary of art
Descriptors
A must-learn
In Midjourney, descriptors are key terms like emotions, time periods, and colors that shape the outcome of generated images. Mastering the use of these descriptors allows you to guide the AI in producing more accurate and varied visual results based on your vision.
Prompt: an illustration of a <time period> astronaut cat
Parameters
These can be used to meaningfully change how a prompt generates something. Add these at the end of a prompt.
Aspect Ratio:
the --ar changes the width-to-height ratio of the image. It is represented by 2 numbers, ex: 5:2
the Chaos parameter:
The --c parameter influences how different the 4 generated images are. High values produce more varied results, and lower values are more similar. The range is from 0-100 with the default being 0.
Character Reference:
You can create images referencing specific characters using –cref URL
- The character weight or –cw parameter sets the strength from 0-100. –cw 0 focuses on the face only, with higher values using face, hair clothing, etc.
- You can maintain a consistent style as well by adding in the –sref parameter, more on this later.
The NO parameter:
This parameter tells midjourney what to exclude.
The Quality parameter:
The –q parameter changes how much detail to give an image. It defaults to 1, but goes up to 2.
The Repeat parameter:
The –r parameter runs multiple jobs at once. Accepts between 2-4 for basic subscribers
Style reference:
Used by calling --sref URL
- You can use more than 1 image as a style reference at a time: –sref URL1 URL2 URL3
- You can also use a random midjourny style with –sref random
Style Weight:
- se --sw to set the severity of a stylization when using an –sref. Accepts values from 0-1000 with 100 as the default
- When using multiple style references, you can assign them different weights --sref URL1::2 URL2::1 URL3::1
Style reference seeds:
- Whenever you do -sref random you will notice there is a number that comes up. You can use that number in future style references to prompt images in that specific style.
- You can blend sref codes together: ie –sref 123 456
- You can blend an sref code with an image -sref 123 URL
- And of course you can add weight to sref codes –sref 123::2 456::1
Stylize:
the –s parameter influences how artistic the image is. Higher values are more artistic but may not adhere to the prompt as strongly. The default is 100 but it accepts 0-1000
Tile:
the –tile parameter creates an image that can be repeated seamlessly to create a pattern for fabrics, wallpapers or banners.
illustration of space and stars --tile
Weird:
The –w parameter determines how unusual something is compared to previous midjourney generations. It accepts 0-3000 and the default is 0.
Vary region
Edit and correct images
In Midjourney, the “Vary Region” feature allows you to keep most of an image static while selecting a specific area to change. You can use this repeatedly to tweak certain sections without altering the rest of the image. This tool becomes even more powerful when combined with the Remix mode, where you can introduce a brand new prompt that modifies the selected area while maintaining consistency with the original image. For best results, use a large selection and be very specific in your description to ensure the changes align with your vision. Additionally, tools like --cref can refine these variations further, and you can combine this with photobashing techniques or image weights to guide the AI in producing a highly customized result.
Prompt: 70's sci-fi illustration of an astronaut cat in the style of Roger Dean, Bruce Pennington, Tim White, Chris Moore, Paul Lehr, Chris Foss, and John Harris, teal, light green, seafoam, cyan
You can reference a specific entity or image with a --cref(more on this later).
In the below example we swap a sponge with a specific dragon image.
You can paste images into your prompt and use them in 2 main ways:
- Using a single image as a reference point with text
- Blend multiple images together with(or without) text.
Image weight
How to weight and bash photos
In Midjourney, the --iw (image weight) option lets you control how much influence an image has compared to your text description, allowing you to decide whether the image or text plays a bigger role in shaping the final result. Photobashing complements this by letting you combine elements from different images into a rough composite of your vision, which you can use as a reference. By adjusting image weight and using photobashed references, you can achieve a more precise and tailored outcome that aligns with your unique concept.
A snake-cat
You can use --iw to adjust the importance of the image vs the text. Higher values mean the image has more impact. It ranges from 0-3 with the default value at 1.
- You can assign weights to each image if using multiple
- Prompt: URL1::2 URL2::0.75
- You can get closer to what you want by photobashing an image first. This means to do a rough overlay of its component parts.
- Imagine you want a specific dragon head on a specific car. You can photobash it and reference it to generate the image.
versions
Midjourney offers different versions of its model, each tailored to create unique styles and outcomes. By choosing a specific version, you can influence the look and feel of the generated images, whether you’re aiming for realism, artistic flair, or something in between. These versions give you control over the final output, allowing you to experiment and find the best fit for your creative vision.
Misc Items:
Whether adding text to graphics, reframing an image, or creating mockups, this section will level up your Midjourney generations.
Adding Text:
You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.
Custom Zoom:
You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.