Sukino's Findings: A Practical Index to AI Roleplay

Finding learning resources for AI roleplaying can be tricky, as most of them are scattered across Reddit threads, Neocities pages, Discord chats, and Rentry notes. It has a lovely Web 1.0, pre-social media vibe to it, with nothing indexed or centralized. There's always something cool buried somewhere you haven't discovered yet. To make things easier, I've compiled this comprehensive, up-to-date index to help you set up a free, modern, and secure AI roleplaying environment and know where to find stuff.

If you have any feedback, wanna talk, make a request, or share something, reach me at [email protected] or send an anonymous message via Marshmallow. You can also send a private message to @sukinocreates on Discord, but please don't assume that I'm your personal tech support. While I don't mind receiving questions that could be added to the index, don't be lazy! Read the page, especially the FAQ section, to see if your question isn't already there.


This index is regularly updated with new links and minor rewrites. These are some of the recent highlights:

2025-07-12 — Added NVIDIA NIM to the Free Providers section.
2025-07-03 — First draft of the reworked If You Want to Run an AI Locally section, now with actual instructions on how to quickly start using KoboldCPP. Consequently, the other backends and broken links were removed.
2025-07-02 — Updated the Free Providers section again to reflect Chutes' new requirement of 5 USD to verify your account and access the free tier. Reworked the Getting Started section to include some basic information about LLMs and how to connect to them. And, hey, there's a new banner for the summer!
2025-06-28 — Updated the Free Providers section to reflect Chutes' new limit of 200 requests per day and tried to make it more readable.



Getting Started

This section will help you quickly set up your roleplaying environment to the point where you can send and receive messages from the default character. First, though, let's establish some important expectations.

These AI models aren't actually artificial intelligences, but LLMs, or Large Language Models, super-smart text prediction tools trained to follow instructions. They can’t think, learn, or be creative; only recombine and replicate the texts in it's own training data.

This means that each model develops its own personality, knowledge gaps, quirks, and biases based on how and what it was trained on. That's why there isn't a single best model that can write about everything and please everyone. Experiment with different models, note what each does well, figure out which ones you like. It's fun to see how each model interprets your characters and scenarios!

Lastly, please never use an AI as a substitute for therapy or an emotional companion, especially if your mental health is not at its best. Don't be fooled by its good writing; it cannot help you work through your problems or make nuanced judgments. These models are designed to be likable assistants who rarely disagree with you for too long, and will inevitably start spewing nonsense to reinforce your behavior and perspective, no matter how harmful or unhealthy they may be.

Picking an Interface

The first thing you'll need is a frontend, the interface where roleplaying happens and your characters live. While it doesn't affect your characters' quality directly, a good frontend gives you the tools to get better responses and build richer worlds.

I only recommend open-source options that are actively maintained, store all data locally, and allow you to use chatbots from anywhere, not just the ones they provide. The following options meet these criteria:

  • Install SillyTavern: Repository · Documentation — The one most people use, and I'll assume you do too. It has all the features you need, and most of the content you'll find is made for it. The interface may seem confusing at first, but you'll quickly get the hang of it.
    • Compatibility: Windows, Linux, Mac, Android, and Docker. iOS users must set up remote access or use an alternative solution.
    • How to Install: This page has a really simple installation guide, with instructions in video and text format. Don't be intimidated by the command prompt; just follow the steps. If you're not tech-savvy and are using a PC or Mac, you can also check out Pinokio, a one-click installer for various AI tools, including SillyTavern. If you need a more in-depth Android guide, check this page.
    • How to Access It on All Your Devices: For this, I recommend Tailscale, a program that creates a secure, private connection between all your devices. With Tailscale, you can host SillyTavern on one device and access it from any other device (including iOS), as long as the host device is turned on and you have an internet connection. All your chats, characters, and settings will be the same no matter which device you use. After installing SillyTavern on your main device, follow the Tailscale section of the official tunneling guide.
  • Or Use an Online Frontend — If you can't install SillyTavern, these alternatives can be used on any device with a modern web browser. But they aren't as feature-rich and can't use the extensions and presets in this index.

Setting Up an AI Model

Now, you just need to connect your frontend to a backend, the AI model that powers your roleplaying setup. How you get access to it depends on how it is published:

  • Open-weights models are publicly available for download. You can run them on your own machine or use independent services who will host them for you. Examples include Deepseek, Mistral, and Llama.
  • Closed-weight models are the ones that corporations keep behind closed doors, hosting them themselves or through select services. You must use them directly from their creators. Examples include GPT, Gemini, and Claude.

Currently, each of these LLMs fits into one of the following two groups:

  • Non-reasoning models are the most common. You send a message, and the model generates a response. Straight up.
  • Reasoning models create an outline of the response before writing it. This helps solve logical problems by breaking them down, and working through them step by step. However, for creative tasks, such as conversation or roleplaying, this extra "thinking" step just makes the responses take longer and cost more on paid services. While you shouldn't ignore reasoning models, you may want to disable this step or configure it for minimal thinking when possible.

Each time you send or regenerate a message in the chat, your frontend sends a request to your backend and waits for it to send a response back. Inside this request is your prompt (instructions, character definitions, previous messages, etc.). Based on it the LLM makes a completion (the text it generates) and sends it back inside the response. There are two ways to configure this communication between the frontend and the backend:

  • Chat Completion formats your request as a sequential exchange of messages between two roles, User and Assistant, just like how ChatGPT works. It's a simple, universal format, and your frontend will handle everything behind the scenes to make it work.
  • Text Completion formats your requests as one single block of text, so the LLM can continue writing at the end of it. It's your job to select the correct Instruct template that the model was trained on before using each model, so the frontend knows how to format this block in the way that the LLM can properly understand your prompt.

Once everything is configured, both options will be transparent to you. You will only see your messages and your characters' responses. All you need to know is that they both exist, which one you are using, and that the settings for one don't apply to the other because they work fundamentally differently.

For the backend to allow this connection to happen, it will provide you with an API Key, a randomly generated password that verifies your identity and the models to which you have access. You don't need a new key for each model, just for each backend.

With the basic understanding of how this works and what you will need to configure, open your SillyTavern page and click on the button on the top bar. You will see the following screen:


This is where you establish a connection with your backend. From the API dropdown menu, you can select either a Chat or a Text Completion connection. In the next sections, I will provide the settings that you will need to copy.

As you can see, SillyTavern is configured by default to connect to the AI Horde. These are models that people host on their own machines and allow others to access for free. However, you can use better models; you just need to follow one of the two sections below. Would you like to run an LLM yourself, or would you prefer to use an online service to host it for you?

If You Want to Run an AI Locally

Currently, there are free or inexpensive online models that outperform anything you can realistically run locally unless you have a computer designed specifically for running AI models. For the average user, it's essentially a trade-off: instead of one large, super-smart model, you'll have a variety of smaller models released almost every day, many of them specifically tuned for roleplaying.

It's uncensored, free, and private. Requires a computer or server with a dedicated GPU or a Mac with an M-series chip.

KoboldCPP will be your backend. It's user-friendly, has all the features you'll need, and is consistently updated. Open the releases page on GitHub and read the notes just after the changelog to know which executable you need to download. No installation is required, everything you need is inside this executable. Move it to a permanent folder where it is easily accessible. When a new version is released, simply overwrite it with the new version.

Currently, the models are available in two formats for domestic use: GGUF and EXL2/EXL3. KoboldCPP uses GGUFs. However, before downloading the models, you need to figure out which models your device can run. To do so, you need to understand these basic concepts:

  • Total VRAM is the amount of memory available on your graphics card, or GPU. This is different from your computer's RAM. If you don't know how much you have or whether you have a dedicated GPU, Google or ask ChatGPT for instructions on how to check your system.
  • Context Length is how many past messages the AI can hold in memory, measured in tokens, parts of a word. The ideal range is between 8192 and 16384 tokens; most models can't use more than that effectively.
  • Models have sizes, calculated in billions of parameters, represented by a number followed by B. Bigger model sizes generally means smarter models, but not necessarily better at roleplaying. So, as a rule of thumb, a 12B model tends to be smarter than an 8B model.
  • Models are shared in various quantizations, or quants. The lower the number, the more compact the model becomes, but less intelligent, too. Recommended quant for creative tasks is IQ4_XS (or Q4_K_S if there isn't one available).

Depending on how much VRAM you have, here are the configurations I recommend you start experimenting with using a GGUF at IQ4_XS:

  • 6GB: up to 7B models with 12K context length.
  • 8GB: up to 8B models with 16K context length.
  • 12GB: up to 12B models with 16K context length.
  • 16GB: up to 15B models with 16K context length.
  • 24GB: up to 24B models with 16K context length.

These recommendations are just a rule of thumb to give you a good performance. You can check out this calculator if you want to find what combinations of model sizes, context length and quants you can expect to run.
Actually, you can get away with better models than I'm telling you. Overflowing your VRAM won't necessarily stop things from working; you'll just start sacrificing generation speed. If you want to experiment later, try to find the smartest model that your device can run with an adequate context length before it becomes too annoying.

With this information, go to the open-weight models recommendation section and open the GGUF link of the model you want to try. Open the Files and versions tab, click on the file ending in IQ4_XS.gguf to download it, and then move it to a permanent location. The same folder as the executable is fine.

Now, avoiding running GPU intensive like games, 3D rendering or animated wallpapers, you need as much free VRAM as possible. Run the executable and wait for the KoboldCPP window to appear, and set:

  • Presets to Use CuBLAS if you have an NVIDIA GPU, or Use Vulkan otherwise. If you got the ROCm version, select Use hipBLAS instead.
  • GGUF Text Model to the path of your downloaded model. Just click on Browse and open the GGUF file.
  • Context Size to the desired length of the context window.
  • GPU Layers you can leave at -1 to let the program detect how much of the model it should load into your VRAM. You can set it to 99 instead to make sure it runs as fast as possible, if you have the appropriate amount of VRAM to fully load the model and context.
  • Launch Browser can be unchecked so that it doesn't open a tab with KoboldCPP's own UI every time you run your model.
  • Click on Save Config and save your settings along with your model so you can load it instead of reconfiguring everything next time.

Now, just click on the Launch button and watch the command prompt load your model. If all goes well, you should see Please connect to custom endpoint at http://localhost:5001 at the bottom of the window. Back on SillyTavern, set it up as follows:

  • Text Completion: On the Connection Profile tab, set the API to Text Completion, API Type to KoboldCpp and API URL to the endpoint shown in the command prompt. Check the Derive context size from backend box to ensure it uses the full context length you configured. Click on Connect.

If the circle below turns green and displays your model's name, then everything is working properly. Send a message to the default character and you should see KoboldCPP generate a response.

If You Want to Use an Online AI

Censorship and privacy can be issues here, as these services can log your activity, block your requests, or ban you at any time. Ultimately, though, your prompts are just one of millions, and no one cares about your roleplay. Just stay safe, never send them any personal or sensitive information, and use burner accounts and a VPN if your activity could get you into trouble if tied to you.

Don't mind people who say you shouldn't use free services because they collect your data. Any company that trains LLMs values new, organic data for training its models as much as money, if not more. Even paying customers have their data collected. If privacy is important to you, look for providers that only host open-weight models and don't train them. Without an incentive to harvest your data, protecting your privacy becomes their main selling point.

Choose a service to be your backend and pick a suitable preset for the model you will use if you want to help it know how to roleplay and get around any censorship.

Free Providers

Running LLMs is expensive, so free options are usually limited to providers conducting tests or trials with rate limits. Please, don't abuse these services, or they might be taken away, and don't harass the providers when they stop offering them for free.

Currently, we have a few high-quality options with generous quotas:

  • Gemini on Google AI Studio: Get an API Key · About the Rate Limits
    • Rate Limit: 100 requests/day for the flagship Pro model, and 250 requests/day for the Flash model.
      • If you need more requests, go to the Google Cloud Console, click on the name of your project in the top bar, and click New Project to create new ones. Then, return to the AI Studio API Keys page and create a new API Key for each project. Switch between them as each one reaches its limit.
    • Recommended Models: gemini-2.5-pro, and gemini-2.5-flash are the flagship reasoning models. Responses are subject to security and safety checks and are cut off if flagged. Read this guide to get around the filters.
    • Privacy: Requires a Google account, so everything will be tied to your name. If you're not in the UK, Switzerland, or the EEA, your prompts will be collected and used for training purposes. Well, it's Google; you can't expect much else.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion and the Source to Google AI Studio. Enter your Google AI Studio API Key and pick the model you want to use from the Google Model dropdown menu. Click Connect.
      • For other frontends: Their OpenAI compatible endpoint is https://generativelanguage.googleapis.com/v1beta/openai/chat/completions
  • Deepseek and Qwen on NVIDIA NIM: Get an API Key · All Free Models
    • Note: Requires phone number verification via SMS. Unavailable in some countries.
    • Rate Limit: 40 requests/minute. Shared across all models. Traffic from other users may cause throttling.
    • Recommended Models: deepseek-ai/deepseek-r1-0528 and qwen/qwen3-235b-a22b are the flagship reasoning models.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion, the Chat Completion Source to Custom (OpenAI-Compatible), and the Custom Endpoint (Base URL) to https://integrate.api.nvidia.com/v1. Enter your Custom API Key and pick the model you want to use from the Available Models dropdown menu. Click Connect.
  • Deepseek on OpenRouter: Create an Account · Get an API Key · Enable Training in the Free Models section · About the Rate Limits · Your Balance · All Free Models
    • Note: OpenRouter doesn't host models, it just redirects your requests to third-party providers. Free models are only accessible when providers are available.
    • Rate Limit: 50 requests/day. Add a total of $10 in balance to your account once to upgrade to 1,000 requests/day permanently. Shared across all models tagged :free.
    • Recommended Models: deepseek/deepseek-r1-0528:free is the flagship reasoning model, and deepseek/deepseek-chat-v3-0324:free the non-reasoning one.
    • Privacy: Requires opting into data training, but whether your data will be harvested depends on the provider offering the free version. Accepts payment in cryptocurrency if you want to upgrade your account.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion and the Chat Completion Source to OpenRouter. Enter your OpenRouter API Key and pick the model you want to use from the OpenRouter Model dropdown menu. Click Connect.
      • Text Completion: On the Connection Profile tab, set the API to Text Completion and the Chat Completion Source to OpenRouter. Enter your OpenRouter API Key and pick the model you want to use from the OpenRouter Model dropdown menu. Click Connect.
      • For other frontends: Their OpenAI compatible endpoint is https://openrouter.ai/api/v1/chat/completions.
  • Deepseek on Chutes: Create an Account · Get an API Key · About the Rate Limits · Your Balance and Quotas · All Free Models
    • Note: Since July 2025, there is a one-time verification fee of $5. Whether or not you consider this "free" is up to you.
    • Rate Limit: 200 requests/day. Message regenerations/reroll consume only 1/10 of a request. Shared across all models.
    • Recommended Models: deepseek-ai/DeepSeek-R1-0528 is the flagship reasoning model, and deepseek-ai/DeepSeek-V3-0324 the non-reasoning one.
    • Privacy: Doesn't require any identification to register. Accepts payment in cryptocurrency if you want to verify your account.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion, the Chat Completion Source to Custom (OpenAI-Compatible), and the Custom Endpoint (Base URL) to https://llm.chutes.ai/v1. Enter your Custom API Key and pick the model you want to use from the Available Models dropdown menu. Click Connect. Note that you will lose access to some samplers using this option.
      • Text Completion: On the Connection Profile tab, set the API to Text Completion, the API Type to vLLM, and the API URL to https://llm.chutes.ai/. Enter your vLLM API key and pick the model you want to use from the vLLM Model dropdown menu. Click Connect.
      • OpenRouter Integration: If you prefer to simply continue using your OpenRouter connection, follow the steps to set it up, then open the Integrations page, click on Chutes and insert your Chutes API Key. On SillyTavern, select only Chutes as your Model Providers on your connection settings, and leave Allow fallback providers checked.

We also have a few providers with smaller models. Keep them in mind in case the larger models become unavailable, you reach your daily quota, or you want to try different models:

  • Mistral on Le Plateforme: API Key · Rate Limits
    • Note: Requires phone number verification via SMS and opting into data training.
    • Rate Limit: 1,000,000,000 tokens/month for each model.
    • Recommended Models: mistral-large-2411 is the flagship model, but mistral-small-2506 is a decent, more up-to-date option.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion, and the Chat Completion Source to MistralAI. Enter your MistralAI API Key and pick the model you want to use from the MistralAI Model dropdown menu. Click Connect.
  • Command on Cohere: API Key · Rate Limits
    • Rate Limit: 1,000 requests/month for each model.
    • Recommended Models: command-a-0325 is the flagship model, and command-r-plus (not 08-2024) is a decent, older option.
    • How to Connect:
      • Chat Completion: On the Connection Profile tab, set the API to Chat Completion, and the Chat Completion Source to Cohere. Enter your Cohere API Key and pick the model you want to use from the Cohere Model dropdown menu. Click Connect.
  • Free LLM API Resources — Consistently updated list of revendors offering access to free models via API.
  • KoboldAI Colab: Official · Unnoficial — You can borrow a GPU for a few hours to run KoboldCPP at Google Colab. It's easier than it sounds, just fill in the fields with the desired GGUF model link and context size, and run. They are usually good enough to handle small models, from 8B to 12B, and sometimes even 24B if you're lucky and get a big GPU. Check the section on where to find local models to get an idea of what are the good models.
  • AI Horde: Official Page · FAQ — A crowdsourced solution that allows users to host models on their systems for anyone to use. The selection of models depends on what people are hosting at the time. It's free, but there are queues, and people hosting models get priority. By default, the host can't see your prompts, but the client is open source, so they could theoretically modify it to see and store them, though no identifying information (like your ID or IP) would be available to tie them back to you. Read their FAQ to be aware of any real risks.

Most of these options operate using a pay-per-request model. You need to pay for both the input tokens, which represent the amount of text the AI needs to read, and the output tokens, which represent the length of the AI's response. So, the longer your session is, the more expensive it becomes.

To prevent the cost of your sessions from skyrocketing, use a provider with Context/Prompt Caching for expensive models. The provider essentially saves your last request, and if your bot's definitions or previous messages haven't changed, you get a cache hit and a discount, since the AI doesn't have to reprocess those parts. Read your provider's documentation to learn how their implementation works.

  • Corporate Models:
  • OpenRouter: Models Available · Prompt Caching — You can use this as an intermediary between you and the AI providers to centralize everything and enable you to use a single balance across different AI models. Compare their pricing with that of the official APIs. If you plan to use an expensive model, find out which providers support prompt caching. Bear in mind that you are adding an additional point of failure to your setup, and that models served via OpenRouter may behave differently to those offered directly by the providers.
  • Other Revendors: There are services that provide access to open-weights models at every price point, including subscription-based options. Comparing and listing them is beyond the scope of this index, but here are a few resources to help you find them. These are not the only options, so be sure to do your own research as well.
  • /aicg/ meta — Comparison of how the different services/models perform in roleplay. Don't take this as gospel, they vary depending on the preset and bots you use, but it can help you set your expectations for what you can pay for.

Note that you are free to switch between AIs during a roleplaying session!
So even if you reach the limits of these APIs or they become too expensive, you can simply use another model for a while. Configure a Connection Profile for each AI with your favorite preset and make switching between them a breeze. Check my guide about it.


Where to Find Stuff

Chatbots/Character Cards

Chatbots, or simply bots, are shared in image files, and rarely in json files, called character cards. The chatbot's definitions are embedded in the image's metadata, so never resize or convert it to another format, or it will become a simple image. Just import the character card into your roleplaying frontend and the bot will be configured automatically.

  • Chub AI — The main hub for sharing chatbots, formerly known as CharacterHub. While it's completely uncensored, most bots are hidden from non-registered users. The platform is also flooded with low-quality bots, so it can be difficult to find good ones without knowing who the good creators are. So, for a better experience, create an account, block any tags that make you uncomfortable, and follow creators whose chatbots you like.
    • Chub Deslopfier — Browser script that tries to detect and hide extremely low quality cards.
  • Chatbots Webring — A webring in 2025? Cool! Automated index of bots from multiple creators directly from their personal pages. Can be a great way to find interesting characters without drowning in pages of low-effort sexbots on Chub. I mean, if the creator went to the trouble of setting up a website to host their bots, they must be into something, right?
  • Anchorhold — Automatically updated directory of bots shared on 4chan's /aicg/ threads without the need to access 4chan at all, what a blessing.
  • Character Archive — Archived and mirrored cards from multiple sources. Can't find a bot you had or that was deleted? Look here.
  • WyvernChat — A strictly moderated bot repository that is gaining popularity.
  • Character Tavern — Community-driven platform dedicated to creating and sharing AI Roleplay Character Cards.
  • AI Character Cards — Promises higher-quality cards though stricter moderation.
  • RisuRealm Standalone — Bots shared through the RisuRealm from RisuAI.
  • JannyAI — Archives of bots ripped from JanitorAI.
  • PygmalionAI — Pygmalion isn't as big on the scene anymore, but they still host bots.
  • Chatlog Scraper — Want to read random people's funny/cool interactions with their bots? This site tries to scrape and catalog them.
Character Generators

Nothing beats a handmade chatbot, but it's handy to have the AI generate characters for you, perhaps to use as a base, or to quickly roleplay with an existing character.

Getting Your Characters Out of JanitorAI

If you are a migrating user, and want to take your bots out with you, these may be of interest to you.

Local LLMs/Open-Weights Models

HuggingFace is where you actually download models from, but browsing through it is not very helpful if you don't know what to look for. Here are some of the most commonly recommended models by 2025-07. They aren't necessarily the freshest or my favorites, but they're reliable and versatile enough to handle different scenarios.

When you are ready, you can check out these pages to find more models:

  • Baratan's Language Model Creative Writing Scoring Index — Models scored based on compliance, comprehension, coherence, creativity and realism.
  • CrackedPepper's LLM Compare · Notion Model List — Models classified by roleplay style, their strengths and weaknesses, and their horniness and positivity bias.
  • HobbyAnon's LLM Recommendations — Curated list of models of multiple sizes and instruct templates.
  • Lawliot's Local LLM Testing (for AMD GPUs) — Models tested on an RX6600, a card with 8GB VRAM, valuable even for people with other GPUs, since they list each models' strengths and weaknesses.
  • HibikiAss' KCCP Colab Models Review — Good list, my only advice would be to ignore the 13B and 11B categories as they are obsolete models.
  • EQ-Bench Creative Writing Leaderboard — Emotional intelligence benchmarks for LLMs.
  • UGI Leaderboard — Uncensored General Intelligence. A benchmark measuring both willingness to answer and accuracy in fact-based contentious questions.
  • SillyTavernAI Subreddit — Want to find what models people are using lately? Do not start a new thread asking for them. Check the weekly Best Models/API Discussion, including the last few weeks, to see what people are testing and recommending. If you want to ask for a suggestion in the thread, say how much VRAM and RAM you have available, or the provider you want to use, and what your expectations are.
  • Bartowski · mradermacher — These profiles consistently release GGUF quants for almost every notable new model. It's worth checking them out to see what's new, even if you don't use GGUF models.

Presets, Prompts and Jailbreaks

Always use a good preset. They are also called prompts or jailbreaks, although this name can be misleading because they are not only for making these AI models write smut and violence. The NSFW part is usually optional.

LLMs are corporate-made assistants first and foremost, so providing them with well-structured instructions on how to roleplay and what users generally expect from a roleplaying session will greatly improve your experience. Each preset plays a little differently based on the creator's preferences and the quirks they discovered in the models. Try different presets to find one you like.

Presets for Text Completion Models

Below is a list of presets for Text Completion connections, along with the Instructs with which they are compatible. The instruct template used by your models can usually be found on its HuggingFace page.

How to Use: Click the A button in the top bar to open the Advanced Formatting window, then click Master Import in the top right. Select the new Context Template, Instruct Template and System Prompt from the dropdowns. Read the descriptions to ensure that you don't need to adjust any other settings.

  • sphiratrioth666 — Alpaca, ChatML, Llama, Metharme/Pygmalion, Mistral
  • MarinaraSpaghetti — ChatML, Mistral
  • Virt-io — Alpaca, ChatML, Command R, Llama, Mistral
  • debased-ai — Gemma, Llama
  • Sukino — ChatML, Deepseek, Gemma, Llama, Metharme/Pygmalion, Mistral
  • The Inception — Llama, Metharme/Pygmalion, Qwen — This one is pretty big, so I wouldn't recommend for small models. Make sure your model is smart enough to handle it.
  • CommandRP — Command R/R+
Presets for Chat Completion Models

Unlike Text Completion presets, this format is much more model agnostic. You can pick any of them and they will probably work fine, but they are almost always designed to deal with the quirks of specific models and to get the best experience out of them. So while it's recommended that you pick one that's appropriate for the model of your choice, feel free to shop around and experiment, or test your favorite preset on the "wrong" models.

One thing that always confuses people is Advanced Formatting, the button with the big A on SillyTavern. The Context Template, Instruct Template and System Prompt here only apply to Text Completion users, as Chat Completion doesn't deal with these things, only with roles.

How to Use: Click the slider icon in the top bar to open the Chat Completion Presets window. If the window has a different title, reconnect via Chat Completion. Click Import presetin the top right, and select the downloaded preset from the dropdown. Read the descriptions to ensure that you don't need to adjust any other settings.

You will see these pages talking about Latte from time to time, it is just a nickname for GPT Latest.

More Prompts

These are really good prompts that you need to build or configure yourself, unlike the common presets they aren't ready to import files.

And few weird prompts until I find a place for them:

Sampler Settings

Each time the AI writes a response, it makes predictions about which words in its vocabulary are most likely to produce the sentences that match your prompts.

Samplers are the settings that manipulate how the AI makes these predictions, and have a big impact on how creative, repetitive, and coherent it will be. Learning how to sample effectively is one of the biggest improvements you can make to your roleplaying sessions.

How To Configure Your Samplers
Token/String Bans and Logit Bias

Want the model to stop writing certain words or phrases? This is how to do it.

String Bans is the cleanest option because it acts as a filter on the LLM's output, not a sampler. It backtracks as soon as banned text appears and generates new text. However, as far as I know, only KoboldCPP and exllamav2, which is used by backends like TabbyAPI, support it.

Token Bans and Logit Bias are true samplers, and supported by virtually every backend. Rather than targeting words or phrases, they target tokens. Specifically, Logit Bias allows you to adjust the likelihood that the LLM will generate a token from 0, which has no effect, to -100, which removes the token entirely from the vocabulary. However, since different words can share the same tokens within a model's vocabulary, this can lead to unintended bans.

These are ready-to-import lists to help you deal with the AI slop:

SillyTavern Resources

Extensions
  • Anchorhold Search — In-app search for bots indexed by the Anchorhold.
  • Notebook — Adds a place to store your notes. Supports rich text formatting.
  • Prompt Inspector — Adds an option to inspect and edit output prompts before sending them to the server.
  • Multi-Provider API Key Switcher — Manage and automatically rotate/remove multiple API keys for various AI providers in SillyTavern. Handles rate limits, depleted credits, and invalid keys.
  • EmojiPicker— Adds a button to quickly insert emojis into a chat message.
  • Chat Top Info Bar — Adds a top bar to the chat window with shortcuts to quick actions.
  • Input History — Adds buttons and shortcuts in the input box to go through your last inputs and /commands.
  • Quick Persona — Adds a dropdown menu for selecting user personas from the chat bar.
  • More Flexible Continues — More flexibility for continues.
  • Rewrite — Dynamically rewrite, shorten, or expand selected text within messages.
  • Dialogue Colorizer — Automatically color quoted text for character and user persona dialogue.
  • Greetings Placeholder — Adds dynamic, customizable elements in character greetings.
  • Timelines — Timeline based navigation of chat histories
  • WorldInfoDrawer — Alternative UI for World Info/Lorebooks.
  • SimpleQRBarToggle — Adds a button to toggle your Quick Replies bar.
  • QuickRepliesDrawer — Alternative UI for Quick Replies.
  • QuickReply Switch — Easily toggle global and chat-specific QuickReply sets.
  • Guided Generations — Modular, context-aware tools for shaping, refining, and guiding AI responses—ideal for roleplay, story, and character-driven chats.
  • Stepped Thinking — Forces your AI to generate a character's thoughts (emotions, plans - whatever you wish) before running the regular prompt generation.
  • Tracker — Customizable tracking feature to monitor character interactions and story elements.
  • Message Summarize — This extension reworks how memory is stored by summarizing each message individually, rather than all at once.
  • NoAss — Sends the entire context as a single User message, avoiding the User/Assistant switch, which is designed for problem solving, not roleplaying. Some AIs seem to work better with this workaround.
  • Cache Refresh — Automatically keeps your AI's cache "warm" by sending periodic, minimal requests. While designed primarily for Claude Sonnet, it works with other models as well. By preventing cache expiration, you can significantly reduce API costs..
  • LALib — Library of helpful STScript commands.
  • Character Tag Manager — Centralized interface for managing tags, folders, and metadata for characters and groups.
  • NemoPresetExt — Helps organize your SillyTavern prompts. It makes long lists of prompts easier to manage by grouping them into collapsible sections and adding a search bar.
  • ReMemory — A memory management extension.
  • StatSuite — Basic character state management.
  • Snapshot — Takes a snapshot of the current chat and makes an image of it for easy sharing.
  • Swipe Model Roulette — Automatically (and silently) switches between different connection profiles when you swipe, giving you more varied responses. Each swipe uses a random connection profile based on the weights you set.
  • Repositories:
Themes
Quick Replies
  • CharacterProvider's Quick Replies — Quick Replies with pre-made prompts, a great way to pace your story. You can stop and focus on a dialog with a certain character, or request a short visual/sensory information.
  • Guided Generations — Check the extension version instead. It's more up to date.
Setups
  • Fake LINE — Transform your setup into an immersive LINE messenger clone to chat with your bots.
  • Proper Adventure Gaming With LLMs — AI Dungeon-like text-adventure setup, great if you are interested more on adventure scenarios than interacting with individual characters.
  • Disco Elysium Skill Lorebook — Automatically and manually triggered skill checks with the personalities of Disco Elysium.
  • SX-3: Character Cards Environment — A complex modular system to generate starting messages, swap scenarios, clothes, weather and additional roleplay conditions, using only vanilla SillyTavern.
  • Stomach Statbox Prompts — A well though-out system that uses statboxes and lorebooks to keep track of the status of your character's... stomach? Hmm, sure... Cool.

More Information About Models


How To Roleplay

Basic Knowledge

  • Local LLM Glossary — First we have to make sure that we are all speaking the same language, right?

How Everything Works and How to Solve Problems

The following are guides that will teach you how to roleplay, how things really work, and give you tips on how to make your sessions better. If you are more interested in learning how to make your own bots, skip to the next section and come back when you want to learn more.

  • Sukino's Guides & Tips for AI Roleplay — Shameless self-promotion here. This page isn't really a structured guide, but a collection of tips and best practices related to AI roleplaying that you can read at your own pace.
  • onrms — A novice-to-advanced guide that presents key concepts and explains how to interact with AI bots.
  • Geechan's Anti-Impersonation Guide — Simple, concise guide on how to troubleshoot model impersonation issues, going step by step from the most likely culprit to least likely culprit.
  • Statuo's Guide to Getting More Out of Your Bot Chats — Statuo has been on the scene for a long while, and he still updates this guide. Really good information about different areas of AI Roleplaying.
  • How 2 Claude — Interested in taking a peek behind the curtain? In how all this AI roleplaying wizardry really works? How to fix your annoyances? Then read this! It applies to all AI models, despite the name.
  • SillyTavern Docs — Not sure how something works? Don't know what an option is for? Read the docs!

How to Make Chatbots

Botmaking is pretty free-form, almost anything you write will work, and everyone does it a little differently, so don't think you need to follow templates or formats to make good bots, plain text is more than fine...

  • Character Creation Guide (+JED Template) — ...That said, in my opinion, the JED+ template is great for beginners, a nice set of training wheels. It helps you get your character started by simply filling a character sheet, while remaining flexible enough to accommodate almost any single character concept. Some advice in the guide seems a bit odd, especially on how to write an intro and the premise stuff, but the template itself is good, and you'll find different perspectives from other botmakers in the following guides.
  • Online Editors: SrJuggernaut · Desune · Agnastic — You should keep an online editor in your toolbox too, to quick edit or read a card, independent of your frontend.
  • Writing Resources - AI Dynamic Storytelling Wiki — Seriously, this isn't directly about chatbots, but we can all benefit from improving our writing skills. This wiki is a whole other rabbit hole, so don't check it out right away, just keep it in mind. Once you're comfortable with the basics of botmaking, come back and dive in.
  • Tagging & You: A Guide to Tagging Your Bots on Chub AI — You want to publish your bot on Chub? Read the guide written by one of the moderators on how to tag it correctly. Don't make the moderator's life harder, tag your stuff correctly so people can find it easier.

Now that the basic tools are covered, these are great resources for further reading.

These are guides made with focus on JanitorAI, but the concepts are the same, o you can get some good knowledge out of them too.

Getting to Know the Other Templates

Again, don't think you need to use these formats to make good bots, they have their use cases, but plain text is more than fine these days. However, even if you don't plan to use them, these guides are still worth reading, as the people who write them have valuable insights into how to make your bots better.


Image Generation

W.I.P.

I like to think of this part as an extension of the Botmaking section, since the card's art is one of the most crucial elements of your bot. Your bot will be displayed among many others, so an eye-catching and appropriate image that communicates what your bot is all about is as important as a book cover. But since this information is useful for all users, not just botmakers, it deserves a section of its own.

Guides

Models

Currently, there are three main SDXL-based models competing for the anime aesthetic crowd. This is a list of these base models and some recommendations of merges for each branch:

Resources

  • Danbooru Tags: Tag Groups · Related Tags — Most anime models are trained based on Danbooru tags. You can simply consult their wiki to find the right tags to prompt the concepts you want.
  • Danbooru/e621 Artists' Styles and Characters in NoobAI-XL — Catalog of artists and characters in NoobAI-XL's training data, with sample images showing their distinctive styles and how to prompt them. Even if you're using a different model, this is still a valuable page, since most anime models share many of the same artists in their training data.
  • Danbooru Tag Scraper — More updated list of Danbooru tags for you to import into your UI's autocomplete. Also has a Python script for you to scrape it yourself.
  • AIBooru — 4chan's repository of AI generated images. Many of them have their model, prompts and settings listed, so you can learn a bit more of many user's preferences and how to prompt something you like.

FAQ

What About JanitorAI? And Subscription Services with AI Characters? Aren’t They Good?

I start the index by sharing my thoughts on what makes a good frontend. Aside from failing to meet those requirements, I don't recommend them on principle.

This hobby is fairly new, and all of these tools were developed through the collaborative efforts of a community of hobbyists who share knowledge and build on each other's work. These people freely distribute code, chatbots, and presets so that others can use, modify, and republish them for everyone's benefit.

JAI lacks many basic features and is built on the open-source code of another site that has been specifically modified to hide the chatbot definitions and lock users into its ecosystem. And those paid services are usually even worse, just leeching off these community-developed resources to make money while contributing nothing back.

If you use one of these services and are happy with it, then by all means, continue using it. I just have strong opinions about solutions that exploit or exclude the rest of the community, and I won't support or promote their use.

What Is the Best DeepSeek Version?

First, please remember that there is no single best LLM. Read the introduction to the Getting Started section if you're not sure why. That said, the best one for you will depend on what you want and what your chatbot needs. In my opinion:

  • The new R1-0528 is the most versatile of them. It performs well with anything I throw at it, but almost too well. It takes any trait you give your characters to the extreme, to the point that intelligent characters become difficult to read and sound fake. You may need to experiment with how to create a well-balanced character.
  • The original R1 is much more unhinged and creative. It pulls things out of thin air. But it tends to overthink and blow every small detail out of proportion. R1's craziness makes it shine in scenarios that require creativity, such as surreal/nightmarish, mystery, or absurd premises.
  • The non-reasoning V3-0324 is way more stable and grounded, but a bit repetitive (and loves to use asterisks everywhere, for some reason). It's better suited to more mundane bots, like realistic, low-stakes or slice-of-life scenarios.

Besides the big three, you will see some other models with the name Deepseek around:

  • The original V3 is simply an inferior version of the V3-0324.
  • R1T-Chimera is an weird mix of V3-0324 and the original R1, created by a third party. This resulted in a model that amplified the flaws more than the strengths.
  • MAI-DS-R1 is a version of the original R1 retrained by Microsoft. It's more stable, but the model has been censored. I don't see any reason to use this instead of the new R1-0528.
  • To be fair, I don't even know what R1-Zero is supposed to be. It feels like a worse original R1, though.
  • Coder models are for programmers, terrible for creative tasks.
  • The ones ending in distill are part of a series of smaller models that were retrained using Deepseek's responses to make them act like the original Deepseek. However, it doesn't work very well. They're fake Deepseeks and really bad for creative tasks.
Why Is the AI’s Reasoning Mixed in with the Actual Responses?

This reasoning step shouldn't be visible to you unless you open the Thinking... window above the model's turn. If they aren't automatically separated, you probably need to change the Reasoning Formatting for your model.

In SillyTavern, to find this option, click on the AI Response Formatting button, the third one with an A in the top bar, and expand the Reasoning section to enable the Auto-Parse option and select the correct Reasoning Formatting. If you are unsure what to write, edit your model's last response to see the prefix and suffix it uses to enclose the thinking step. It's usually something like: <think></think> or <thinking></thinking>

How Do I Make the AI Stop Acting for Me?
  • Make sure you are using a preset that gives the AI rules that reinforce that your persona is yours and yours alone to control.
  • Check if the bot's example dialogues and greetings do not show the AI acting on your behalf. Read them and consider whether you would be okay with the AI responding to you exactly as they are written. If not, change them.
  • Maybe you aren't giving enough information for the AI to work with, so it takes over your character to push the narrative forward and meet the expected response length.

I have two guides that can help you figure this out: Make the Most of Your Turn; Low Effort Goes In, Slop Goes Out!, it even has an example session of how I roleplay, and The AI Wrote Something You Don't Like? Get Rid of It. NOW! Also check Geechan's Anti-Impersonation Guide and Statuo's section on this problem, where he explains other possible causes and rants about the nature of AIs. With these guides, you should have a good understanding of why it's happening and how to make it stop. Yes, you need to read up on how to roleplay effectively and which bad practices cause it. There is no magic bullet.

How Detailed Should My Persona Be?

Keep it minimal, only include information that you want the AI to use. Try imagining that you are an outside observer of your character. Describe their appearance and how other characters perceive them. Maybe include minor world-building details, such as rumors about your character and their reputation. You don't need to provide an in-depth description of your persona's past, personality, or inner workings. You'll be playing the character, not the AI.

I Got a Warning Message During Roleplay. Am I in Trouble?

Probably not. You just received what we call a refusal, a safeguard that the model's creators incorporated into the training data to prevent the model from writing about certain topics. LLMs can only generate text; they can't analyze or report on your activity on their own.

Those who run the LLM on their own machines or use privacy-respecting services don't have anything to worry about. Simply rewrite your prompt to try to get around the refusal, or look for a less censored model.

However, if you use an online API that logs your activity, the people behind it may use external tools to analyze your logs, and take action if they see too many refusals or notice that you are prompting their models to generate content about controversial or illegal topics.

In any case, if you're in real trouble, it won't be the AI model that informs you. Instead, you'll receive warnings on the provider's dashboard or via email, or you'll be banned directly.

My Provider/Backend Isn’t Available via Chat Completion. How Can I Add It?

Check their pages and documentation for an OpenAI-compatible endpoint address, which looks like this https://api.provider.ai/v1. Basically, it mimics the way OpenAI's ChatGPT connects, adding compatibility with almost any program that supports GPT itself.

To use it, create a new Chat Completion connection with Custom (OpenAI-compatible) as the source, and manually enter the Custom Endpoint address and your API key in the appropriate fields. If the model list loads when you press the Connect button, you are golden, just select the right model there.


Other Indexes

More people sharing collections of stuff. Just pay attention to when these guides and resources were created and last updated, they may be outdated or contain outdated practices. A lot of these guides come from a time when AI roleplaying was pretty new, we didn't have advanced models with big context windows, and everyone was experimenting with what worked best.


Previous versions archived on Wayback Machine and on archive.today.

Edit
Pub: 08 Feb 2025 03:42 UTC
Edit: 21 Jul 2025 00:31 UTC
Views: 89328