Jun 28, 2026

Updated: Jun 29, 2026 3:19 pm

Civitai Changelog — late May to June 2026

🚀 New Features

• Krea2 — A base model made from scratch by the Krea team, tuned for creative generation; supports LoRAs and a fast turbo mode.
• Boogu — New multimodal image model for generation and editing, with standard + fast Turbo variants.
• HappyHorse 1.1 — New video model: text-to-video, image-to-video, reference-to-video, and video editing.
• MAI — Microsoft's MAI-Image-2.5 image checkpoint with image editing (img2img) and auto-cropping for edits.
• Anima for everyone — The Anima image model, previously members-only, is now available to all users.
• Krea 2 & Boogu LoRA training — Train LoRAs on the new Krea and Boogu models.
• New login experience — Rebuilt login/OAuth with single sign-on across civitai.com and civitai.red.

✨ Quality of Life

• Base-model filter is now a searchable dropdown (like tags).
• Shift-click range select — Select a range of items at once.
• Training setup redesign (Step 3) + a new "Save every" control for intermediate epochs.
• Discord-style timestamps — Use <t:UNIX:STYLE> in markdown; renders in each viewer's local time.
• Posts in nav is now an opt-in setting; scheduled posts get a filter chip.
• Blocking improvements — clearer block screens + Report on blocked-user cards.
• Like/Save buttons now give feedback in the resource-add menu; clearer Blue-Buzz-can't-do-NSFW messaging.
• Faster homepage load and accessibility improvements.

🐛 Bug Fixes

• "Train Further" now trains on your model's original base model instead of defaulting to SDXL.
• Reactions — correct reacted-state highlighting, fixed feed reaction counts, own-reactions no longer echo in live counts.
• Buzz — Claim button no longer clipped to "Clai"; reward writes fail-soft so a hiccup won't break your action.
• Publishing — publish dates preserved on republish/restore (and locked after first publish).
• Reliability — downloads and image search fail gracefully instead of erroring when a backend is slow.
• Comments-loading and chat-history fixes; profile avatar rendering fix.
• Overdue members can now cancel their membership; images re-index on re-publish; monthly membership badges grant correctly.
• Model hashes — Files that were showing an incorrect 603AF6DB hash now display their correct, updated hashes.

🧪 In Testing

Rolling out to a limited group first — coming to everyone soon.

• 3D Models — Generate 3D models with rigged/animated/motion variants (PolyGen), an interactive viewer with animation playback, and dedicated 3D feeds and tools.
• Apps & Marketplace — A developer apps marketplace to build, publish, and share apps that plug into Civitai.
• ControlNet — Temporarily paused while we make improvements; coming back soon.

🌟 Coming Soon

• Custom Anima training — You will be able to use custom Anima checkpoints for training LoRA models.

May 22, 2026

Updated: May 22, 2026 5:54 pm

ANNOUNCEMENT

CHANGELOG

Another batch to share. A new product launching in beta, a couple of long-requested training updates, the generator getting cheaper at scale, and a fresh page for tracking known issues. Plus a stack of small fixes and a sunset note at the end.

All of the following is live now.

🎵 Civitai Radio (Beta)

Civitai Radio is live in beta. Tune into community-built stations playing AI-generated music on demand, or spin up your own station, pick a DJ voice (or record your own), and let people tune in to you. The full article has the details and what's coming next for music on Civitai.

⚙️ Anima Training: Member Early Access

We launched Anima Base generation on-site last week. This week, Members get early access to Anima training. If you've been waiting to fine-tune in this ecosystem, you can now.

🎨 ControlNet Support for SD1, SDXL, Flux.1, and Z-Image

The popular feature from 2023 is finally on-site. You can now generate images guided by the outlines or depth of another image. Find it in the advanced options inside the text-to-image workflow for these ecosystems.

💰 Bulk Discount on Image Generation

Generate more at a time, pay less per image. Starting at 2 images per batch you get a small discount, scaling up to 25% off at 12 images. Applies to: SD 1.5, SDXL, Pony, Illustrious, Flux.1, Flux.2 Klein, Z-Image, HiDream, HiDream o1, and Anima. The discount comes off the generation cost; license fees (like Anima's) are unchanged.

🎯 Known Issues Page

I get ragged on every time we post about new features. "What about the bugs?" Fair. So we built a Known Issues page. Active bugs we're working on, things we believe are fixed and are reviewing, and the rest. This is a brand-new page and we'll be refining it as we use it. Go check it out next time you're wondering whether something is on our radar.

🛠️ Quality of Life

• Better empty state on post pages: when images on a post are hidden by your filters, you now get a contextual filter notice explaining why, instead of the generic "no images" message.

• Moderation flags on the post editor: image cards in the post editor now surface needs-review and minor status so you can spot moderation state at a glance. The publish confirmation and review warning dialogs were also consolidated so you don't get prompted twice.

• Scheduled images filter: there's now a "Scheduled" chip in the media filter dropdown, default-off. Previously your own scheduled images were pinned to the top of every feed whether you wanted that or not. Now they're opt-in.

🐛 Bug Fixes

• Followed Only filter no longer resets when you hit Clear Filters inside the filters panel. The Followed scope is preserved.

• GGUF quant downloads now actually download the variant you picked. Previously every variant was falling back to the default quant.

• PG13 content on civitai.red while logged out now shows the sign-in prompt instead of a confusing "no images visible at your current browsing settings" message. Same behavior as civitai.com.

• Person-of-interest tags (Celebrity, Actor, Actress, Real Person, etc.) can no longer be attached at model upload time.

• Autocomplete now says person-of-interest searches are filtered, not blocked. The old wording made it sound harder than it is.

• Account menu items now respect feature-flag visibility correctly when the flag is undefined.

👋 Kohya Trainer End-of-Life

We're retiring the Kohya trainer. We've fully transitioned to AI Toolkit and, after a lot of community feedback and iteration, it's now performing as well or better than Kohya in most cases. Because Kohya is end-of-life, we've also had to apply a slight price increase to the remaining runs.

If you've been on Kohya, this is a good week to try AI Toolkit.

May 20, 2026

Updated: May 22, 2026 3:41 pm

IMAGE & VIDEO FEEDS

We wanted to share what's behind the issue some of you have seen where images or models seem to flip between published and unpublished, or where one person can see something in a feed and another can't.

What's happening

Our image feeds, the main feed, model galleries (the images shown on a model's page), and user image feeds — are powered by a search index called Meilisearch. To keep up with the volume of content on the platform, we run two Meilisearch instances side by side. Each one receives updates from our main app independently, and they don't always process those updates at the same pace. One instance might be a bit behind the other at any given moment.

Until today, requests were round-robined between the two instances. That meant if you refreshed the page, you might hit a different instance each time. If those instances disagreed about the current state of a piece of content, you'd see things appear, disappear, and reappear between refreshes.

What we just shipped

We've changed how requests are routed so that you'll consistently be served by the same instance rather than bouncing between them. This won't fully eliminate the underlying inconsistency — one user may still see something another doesn't until both instances catch up — but it does stop the flicker where content seems to vanish and come back on your own screen.

Your content is safe

This is a routing and indexing issue, not a data issue. Your images, models, and ratings are intact on our end. There's nothing you need to delete or re-upload.

The longer-term fix

We've been building a replacement for this indexing system for several months. It's designed to give us real-time, consistent feed updates across all users — no more minutes-long lag, no more disagreement between instances. We hit some setbacks in our initial implementation and are taking another pass at it, but it's actively in progress. When it lands, feeds won't just be more consistent — they'll be dramatically faster.

Thanks for your patience while we get this across the finish line.

ANNOUNCEMENT

CHANGELOG

A handful of updates to share this round. New ways to tell a story, new ways to get featured, new things in the generator, and a fresh way for community curators to highlight great work.

All of the following is live now.

📚 Civitai Comics

We launched Civitai Comics so you can tell your story page by page. Thousands of comics have already been made. Not everything is published yet, but it's a strong start. The full article has all the details on how it works.

🖼️ Featured Image Requests Have Been Removed

This one is subtle, and most of you probably never noticed it. Underneath the featured images on the homepage, there used to be a way to request that an image be featured. We're not pursuing that flow anymore, so we've taken it out.

If you want to get on the homepage, the most reliable approach is to make great images and post them. The next item below also opens up another path.

🎲 Featured Images: More Variety

Featured images used to be a fairly random pull from the most recent images added to featured images and curated collections. We tightened the logic so a single creator now appears at most twice in any pull, and the order shuffles more aggressively. The result is more variety from more creators on the front page.

🎬 LTXV Sulphur in the Generator

Sulphur dropped last week and is already available in the generator. Try it out.

⚡ LTX Promo: Video From 35 Buzz

LTX video generation now starts at 35 Buzz. That's cheaper than generating an image with some models.

Members can also generate multiple LTXV videos at once:

• Bronze: 2
• Silver: 3
• Gold: 4

🌟 Featured Collections

We quietly rolled this out a little while back without making noise about it: select community collections now get featured on the homepage. That means more curators than ever are looking for ways to surface up-and-coming artists and other great creators, with taste from a variety of perspectives.

If you'd like to be considered as a featured curator, reach out at [email protected].

Big thanks to our first curators: Cinnadust, Snilek_Robotka, vishnarjk, purplelady, Daalis, and Alexia_Rousse.

Apr 30, 2026

Updated: May 08, 2026 5:27 pm

ANNOUNCEMENT

CHANGELOG

The past couple of months have been busy, and at times a little turbulent. Between outages, infrastructure changes, and the .com/.red domain split, we know things haven’t always been as smooth as they should be.

A lot of work has gone into stabilizing the platform. While backend teams have been focused on tackling those issues, improving reliability, and getting everything onto a stronger footing, frontend development hasn’t slowed down. In parallel, we’ve continued shipping improvements, refining existing features, and rolling out new ones.

All of the following updates are live and ready for you to explore.

🖌️ Generator Updates

Over the past couple of months, we’ve introduced a range of meaningful quality-of-life improvements and new features to the Generator, all designed to make the experience faster, more streamlined, and more intuitive.

Generator Metadata Reader

A long-requested addition, now available.

From the Image Generation menu, you can select Extract Metadata and drag any image generated on Civitai into the dropzone.

This allows you to view the full prompt and complete metadata embedded within the image, along with quick access to actions such as Remix and Image-to-Image.

This makes it significantly easier to revisit previous generations, understand how they were created, and build on them. For users working with images downloaded from the site, it provides a simple and reliable way to recover and reuse generation settings.

Prompt Enhancement

A major update, and one that’s been a long time coming.

Accessible via the blue ✨ Enhance link above the Prompt box, you can now use an LLM to expand your prompts, adapt them to different prompting styles, and refine them with richer, more descriptive language, making it easier to get consistent, high-quality results.

Each enhancement costs 1 Buzz (of any color).

Generator filtering

It may feel like a feature that should have always existed, but expanded Generator filtering is now, finally, here.

You can now filter your generations by model ecosystem and workflow (img2img, Draft, Upscale, and more), with fully searchable filters to help you quickly find exactly what you’re looking for;

Generation Presets

Released just this week, Generation Presets take the effort out of restoring prompts and advanced Generator settings.

If you regularly switch between different ecosystems, you can now save your preferred configurations and load them instantly. Simply set up your Generator the way you like, save a preset per ecosystem, and reapply it with a single click.

This removes the need to manually reconfigure settings each time, making it much easier to move between workflows without disruption.

This is just the beginning for Presets. We’re already exploring features like shareable, public presets and additional ways to build on this system!

🖼️ New Image Gen Models

We’ve seen a wave of impressive new image models released over the past few weeks, and have added several of them for on-site generation. Each brings its own strengths, from prompt understanding to text rendering and editing workflows.

Grok Imagine: Built by xAI, Grok’s image generation focuses on fast, flexible text-to-image creation with strong real-world prompt interpretation. It supports both generation and image editing workflows, allowing you to modify existing images as well as create new ones.

One of its defining traits is a more permissive and flexible approach to prompt handling, making it well suited for rapid experimentation and less constrained creative workflows. Not as spicy as it once was - xAI have had to lock it down somewhat due to backlash.

GPT Images 2.0: Developed by OpenAI, GPT Images 2.0 represents the latest generation of OpenAI’s image models - and a huge leap forward from GPT Image 1. It combines strong prompt understanding with ultra-realistic image generation and precise editing capabilities. Compared to earlier models, it introduces much improved reasoning, better instruction following, and more reliable handling of complex prompts and compositions.

Ernie Image: Created by Baidu, Ernie Image is a high-quality text-to-image model with a focus on multilingual support and strong text rendering within images. It handles both English and Chinese prompts effectively and includes built-in prompt expansion capabilities, helping produce more complete and descriptive outputs with less effort. ERNIE Image is an open-weight model released under a permissive license, meaning it can be run locally, modified, and used commercially. It's also great for LoRA training (see below!)

Wan 2.7 Image: A next-generation text-to-image and instruction-based editing model from Alibaba’s Wan series, designed to address long-standing issues in AI image generation such as repetitive faces, unstable text rendering, and poor control over color and layout.

Rather than just improving visual quality, Wan 2.7 focuses on precision and reliability. It introduces stronger prompt understanding, more stable outputs, and fine-grained control over key elements like identity, typography, and composition.

Qwen Image 2: Developed by Alibaba, it combines text-to-image generation and image editing into a single 7B parameter unified model, with strong prompt fidelity, native high-resolution output, and excellent handling of complex scenes. No open weights yet, but we live in hope!

📽️ New Video Gen Models & Features

We’ve also seen a number of exciting advances in video generation over the past few weeks, and have introduced several new models for on-site use. Each brings its own strengths, from prompt understanding and motion consistency to scene coherence and stylized output.

Wan 2.7: A highly flexible, all-in-one video model from Alibaba, supporting text-to-video, image-to-video, and reference-to-video, with an in-built prompt-enhancer (optional), and up to 1080p outputs.

Grok Video + Edit Video: Focused on editing and transformation, allowing you to modify existing video using prompts, at 720p! Well suited for remixing, restyling, and iterating on clips rather than generating from scratch.

Seedance 2.0: Generated significant attention on release, before being temporarily withdrawn over concerns around realistic human depiction. It’s now back, and available for use. Seedance 2.0 is known for strong motion quality and natural, lifelike animation, making it well suited for dynamic and character-driven scenes.

Happy Horse 1.0: From Alibaba, currently sitting at the top of the global video generation leaderboards, HappyHorse 1.0 has quickly established itself as one of the strongest models available. It ranks #1 on the Artificial Analysis Video Arena, a benchmark based on blind user preference testing, outperforming models like Seedance 2.0 by a significant margin. It's pricy, but it produces unparalleled content!

LTX: We’ve expanded LTX’s capabilities with two key additions. LTX 2.0 now supports first-and-last frame control, giving you more precise control over how a scene begins and ends.

In addition, LTX 2.3 now supports full 20-second video outputs, enabling longer, more complete clips without the need to stitch shorter generations together.

⚙️ New LoRA Training Models

We’ve introduced several new LoRA training model options recently, giving you the ability to train within some of the most exciting new text-to-image ecosystems, as well as the powerful LTX 2.3 video model.

Ernie training

Ernie has proven to be a strong and accessible option for LoRA training. In testing, it produces high-quality results from relatively small datasets (around ~30 images), while remaining flexible and easy to work with across a range of styles and subjects.

Training costs for Ernie start at 1,000 Buzz (any color).

LTX 2.3 training

LTX 2.3 represents a major step forward in video LoRA training, delivering exceptional motion-focused LoRAs. Compared to earlier approaches, it produces more consistent and natural results, particularly for movement and scene continuity. It also supports training from datasets that include video clips, allowing for more expressive and dynamic LoRAs.

🔉 New Modality! Audio Gen Coming Soon!

We’re expanding into a brand new modality: Audio Generation.

The upcoming Generator audio panel will introduce Ace-Step Audio powered generation, along with Audio LoRA training, and Civitai Radio, bringing sound into the same creative workflow as images and video.

We'll be sharing more details soon!

💬 Comic Gen Update

A couple of updates ago, we teased a new Comic Creator tool, and after many rounds of testing and iteration it’s now almost ready!

Focused on multi-panel storytelling, it brings character consistency, scene continuity, and structured layouts into a single workflow, opening up a whole new way to create, on Civitai.

We’re excited to get this into your hands and see the stories you bring to life!

Apr 23, 2026

Updated: May 08, 2026 5:31 pm

A New Structure for Models

When Civitai first launched in late 2022, things were simple. A “model” was just a single file - typically a checkpoint (.ckpt) - containing everything needed to generate images.

Not long after, LoRAs changed that.

They introduced a new idea: models didn’t have to be all-in-one files anymore. Instead, they could be smaller, modular pieces that modify or enhance a base model. That was the first step toward a more modular ecosystem.

Since then, things have evolved quickly.

Modern workflows now often rely on multiple components working together rather than a single monolithic file. A typical setup might include a:

• UNet (the core denoising network)
• Text Encoders (which interpret prompts)
• VAEs (which handle image encoding and decoding)
• LoRAs and other adapters layered on top
• Optional Upscalers, Frame Interpolation Models, or Background Removal models

Because of this, the idea of “a model” as a single container has fundamentally changed. It’s no longer always one file. It’s often a collection of interoperable parts that work together to produce a result.

A New Structure for Models

To reflect this shift, we’re evolving how models are structured on Civitai.

A Model is now best understood as a container. It represents a concept, project, or family of related assets.

A Model Version represents a specific, usable artifact or configuration within that model. This could be:

• A full checkpoint
• A LoRA
• A json workflow file
• A UNet
• A VAE
• Any or other component

This structure allows creators to group related work together, even when those assets are very different under the hood. It also makes it easier for users to understand compatibility, and what files are actually required to use a resource they find on Civitai.

What This Means

We’ve moved from a world where:

A model = one file

to one where:

A model = a structured collection of components and versions

This change reflects how modern generative workflows actually work and gives us a much stronger foundation moving forward.

We’ve overhauled the Model Detail sidebar to support this new paradigm, and the Model Upload interface now allows you to upload multiple components or link to existing Civitai models, reducing the need for duplicate uploads.

Migration

We’ve built a process to help you transition from the old system to the new one, if you choose. There’s no need to convert any existing models unless you want to!

From the Model Context Menu, use the "Merge Versions" option.

Note that this is a one-way conversion - once finalized, there's no going back!

We're Listening

This is a pretty big step forward, and we expect there may be a few bumps along the way. If anything feels off, confusing, or just not quite right, let us know! Your feedback helps shape how this evolves, and we’ll be actively refining things based on what we hear from you.

Apr 15, 2026

Updated: Apr 24, 2026 11:08 am

ANNOUNCEMENT

CIVITAI.COM

CIVITAI.GREEN

CIVITAI.RED

Last week, we announced a big shift for Civitai: two front doors. One mainstream-friendly, one freedom-first, so both sides of the community can actually thrive.

That transition is now underway.

• Civitai.com becomes the PG/PG-13 experience (what Civitai Green was)
• Civitai.red becomes the freedom-first experience (what Civitai.com was)
• Civitai.green will redirect to Civitai.com, so existing links and bookmarks keep working

Nothing is being deleted, reset, or moved in the way people fear. Your account, your Buzz, your models, your images, your posts. It all stays exactly as it is. The two domains are simply different views of the same platform.

If you currently use Civitai for everything:

👉 Bookmark civitai.red and you’re done

If you’re here for SFW content or have been using CIvitai.green:

👉 You don’t need to do anything

What to Expect During the Transition

We’ll be actively switching domains, filtering content, and updating systems behind the scenes. That means:

• You may see temporary inconsistencies in content visibility
• Some features may feel slightly out of sync between domains
• Login, generation, content feeds, and other site services may have brief hiccups as things settle

We’ve done a lot to make this smooth, but with a change at this scale, a bit of disruption is unavoidable.

Additional Changes Rolling Out

Alongside the domain split, a few platform updates are going live:

• Direct ads now appear on Civitai.red only
• Buzz selection in the Generator: you can now choose which color of Buzz to spend first, instead of always consuming Blue Buzz by default

We’re also still working toward full feature parity between Civitai.com and Civitai.red. Most core functionality is already aligned, but some differences will exist in the short term as we continue syncing both experiences.

What’s Changing (and What Isn’t)

Changing:

• Civitai.com becomes strictly PG/PG-13
• NSFW content moves to civitai.red
• Credit card payments, memberships, and mainstream features return to .com
• Crypto remains the primary payment method on .red
• The Creator Program will now offer a unified .com/.red pool - and the ability to bank Green and Yellow Buzz

Not changing:

• Your account, data, and content
• Our stance on NSFW content
• The core platform and community features

This is not a reset. It’s a separation that lets both sides grow properly.

Why This Is Happening

Short version: payments and platform reality finally collided.

Mainstream processors won’t work with platforms like ours as long as mature content exists on the same domain.

We tested this with Civitai Green, and it worked.

So instead of keeping Green as an experiment, we’re making it the main front door and giving the rest of Civitai its own space to exist properly, without compromise.

Read the Full Breakdown

If you want the full context, reasoning, and FAQ, you can read the full article here:

👉 https://civitai.com/articles/28369

Apr 01, 2026

Updated: Apr 24, 2026 11:06 am

KONO

ANNOUNCEMENT

We've made several updates to the Knights of New Order rating game to ensure it remains fair and rewarding for honest players:

• Fervor now rewards accuracy – Fervor is now weighted by your accuracy ratio. High-accuracy players will see little change, but low-accuracy players will earn significantly less Fervor.
• Voting speed limits – Limits on how fast and how many ratings you can submit per minute, hour, and day. Vote at a natural pace and you won't notice them.
• Stronger sanity check penalties – Failing sanity checks carries heavier consequences. Severe under-ratings (e.g., rating explicit content as safe) result in immediate penalties. Repeat failures within 24 hours escalate further.
• Automatic account resets – Accounts with suspicious voting patterns (e.g., spamming the same rating for everything) will be automatically reset.
• Vote weight affected by penalties – Smites now reduce your vote weight, meaning penalized players have less influence on rating outcomes until they prove accuracy over time.

These changes are designed to protect the integrity of community moderation while continuing to reward dedicated Knights. For full details, visit our updated Knights of New Order guide.

Mar 26, 2026

Updated: Apr 24, 2026 11:06 am

ANNOUNCEMENT

BUZZ

We’ve made an important update to how Yellow Membership Buzz ⚡ is distributed to address a common pain point and reduce confusion. Instead of Buzz being automatically added to your Yellow Buzz Balance each month, Buzz will now become available for you to claim once it unlocks, at your normal roll-over date. You’ll receive a notification email when it’s ready, and from there you can claim it immediately or leave it to stack up over time.

This change gives you full visibility into your Membership Buzz. You can now clearly see how much you’ve already received, as well as how much is still outstanding - Buzz that has been purchased but is not yet claimed. You’re free to claim it monthly as your Membership progresses or let multiple months build up before claiming all at once. We’ve also added a full history so you can easily track what you've already claimed.

Most importantly, this system helps eliminate issues where Buzz wasn’t properly disbursed. Because Buzz is now claimed manually, it ensures nothing is lost or missed. Any Buzz that hasn’t been claimed won’t expire - it will remain available to you indefinitely once unlocked.

This change only affects Pre-Paid Yellow Buzz Memberships. Direct Yellow Buzz purchases via Coinbase or Crypto Deposit still credit directly to your Yellow Buzz Balance.

You can manage and claim your Buzz at any time from the Membership page or a new section on your Buzz Dashboard.

CRYPTO BUZZ

Mar 17, 2026
Updated: Apr 24, 2026 11:07 am

We’ve expanded the ways you can purchase Buzz with crypto! Alongside our standard Coinbase flow (which can be used to buy Memberships!), you can now deposit cryptocurrency directly to a personal Civitai address, where it will be automatically converted into Buzz and credited to your account!

This new system supports multiple popular cryptocurrencies and networks, making it easier for users around the world to fund their accounts.

To purchase Buzz with crypto using this method - head to https://civitai.com/purchase/buzz - or from the "Buy Buzz" link in the Profile Picture Menu.

How it works

Check out the full Guide to Purchasing Buzz with Crypto - but the basic steps are;

1. Select a supported cryptocurrency.
2. Send funds from any crypto app to your personal Civitai deposit address (make sure you send to the correct address for the selected coin type!)
3. Once the transaction is received, the crypto is automatically converted into Buzz ⚡and credited to your account.

Most transactions complete within about a minute, but can take 15 minutes (and sometimes longer!) to appear depending on the chosen cryptocurrency and network. You can monitor your deposits on the same page;

Key details

• Each user is assigned permanent deposit addresses, so you can reuse it any time. You can even share them on your Civitai profile, for direct tips!
• Deposits work from any wallet or exchange - send to Civitai from PayPal, Cash App, Venmo, Revolut, Moonpay, Binance, Coinbase, and many more!
• Buzz is credited after the transaction is confirmed - this can take some time! Don't panic if your transaction doesn't show up on Civitai immediately!
• Simply send any amount above the minimum deposit for the selected network.

Supported cryptocurrencies

At launch, deposits support a number of commonly used currencies and networks including:

• BTC
• ETH
• DOGE
• LTC
• SOL
• USDT
• USDC
• PYUSD
• SHIB

Additional currencies, features, and improvements to the purchase experience are planned as we continue to refine the system!