How to Use SillyTavern Character Cards with Ollama: A Step-by-Step Guide for 2026

How to Use SillyTavern Character Cards with Ollama: A Step-by-Step Guide for 2026

If you’re looking to run immersive roleplay with AI characters entirely offline, pairing SillyTavern with Ollama is one of the most powerful and privacy-respecting setups you can build in 2026. By combining local LLMs with the rich personality data stored in character cards, you can create dynamic, responsive conversations without relying on cloud APIs. In this guide, we’ll walk through the entire process—from setting up Ollama and SillyTavern to importing character cards and fine-tuning your experience. Along the way, we’ll feature a popular card, Elena the Enchantress, to show you exactly how it works.

---

Why Use SillyTavern with Ollama?

SillyTavern is a frontend that excels at managing character cards—those structured JSON or PNG files that define an AI character’s personality, backstory, and dialogue style. Ollama, on the other hand, is a user-friendly tool for running large language models locally on your machine. When you combine them, you get:

• Full data privacy – All conversations stay on your device.
• No API costs – Run unlimited chats with open-source models.
• Customizable characters – Import or create character cards that define unique personas.
• Offline availability – Roleplay anywhere without an internet connection.

This setup is especially popular among writers, gamers, and hobbyists who want to explore character-driven narratives without censorship or subscription fees.

---

Step 1: Install and Configure Ollama

Ollama supports Windows, macOS, and Linux. Download the latest version from the official website and run the installer.

Basic Setup

1. After installation, open a terminal (Command Prompt on Windows, Terminal on macOS/Linux).
2. Verify Ollama is working by running:
   ollama list
   (You should see an empty list if no models are downloaded yet.)
3. Pull a model suitable for roleplay. For example:
   ollama pull llama3.2:3b
   This model is lightweight and works well on most systems. For better character adherence, try mistral:7b or dolphin-mixtral:8x7b.

Enable Ollama for Remote Access (Optional)

By default, Ollama only listens on localhost. To let SillyTavern connect, you may need to set the environment variable:
OLLAMA_HOST=0.0.0.0
Then restart Ollama. This allows SillyTavern to communicate with it over your network.

---

Step 2: Install and Configure SillyTavern

SillyTavern is a Node.js application. You’ll need Node.js 18+ installed.

1. Clone the repository:
   git clone https://github.com/SillyTavern/SillyTavern.git
2. Navigate to the folder:
   cd SillyTavern
3. Install dependencies:
   npm install
4. Start the server:
   npm start

The terminal will show a local URL (usually http://localhost:8000). Open it in your browser.

Connect to Ollama

1. In SillyTavern’s web interface, click the plug icon (Settings) in the top left.
2. Go to API Connections.
3. Select Ollama from the dropdown.
4. Enter the Ollama API URL (default: http://localhost:11434).
5. Click Connect. You should see a green checkmark.

Now SillyTavern is ready to use your local LLM via Ollama.

---

Step 3: Import Character Cards

Character cards are the heart of your experience. You can find thousands of free cards online, or create your own using a tool like the MiniTavern Character Card Market.

How to Import a Card

1. In SillyTavern, click the character icon (top bar) to open the character manager.
2. Click Import Character.
3. Choose a .png or .json file. PNG cards store both image and data; JSON cards store data only.
4. SillyTavern will parse the card and display the character’s name, description, and example dialogue.

Example: Elena the Enchantress

Let’s import a card for Elena the Enchantress—a wise sorceress with a mysterious past. After import, you’ll see:

• Name: Elena
• Personality: Mysterious, nurturing, cryptic
• Scenario: She guards an ancient library in a hidden forest
• Example dialogue: “The runes speak of your arrival, traveler. What knowledge do you seek?”

This card defines her voice perfectly. When you chat, the local LLM will use this data to stay in character.

---

Step 4: Optimize Your Chat Settings

To get the best roleplay experience, tweak a few parameters in SillyTavern.

Model Selection

In the API Connections panel, choose the model you pulled earlier (e.g., llama3.2:3b). Larger models offer better nuance but require more VRAM.

Response Settings

• Temperature: 0.7–0.9 for creative dialogue; lower for more predictable responses.
• Max tokens: 200–400 for balanced replies.
• Context size: 2048–4096 tokens depending on your model.

Character Card Settings

Under the character manager, you can adjust:

• System prompt – Override the card’s default instructions.
• Jailbreak prompt – Add rules like “Always stay in character” or “Use descriptive narration.”

For Elena, you might add: “Describe the ambient sounds of the library and your flowing robes.”

---

Step 5: Start Roleplaying

Click on Elena’s avatar to begin a new chat. Type your first message, and the local LLM will respond based on her card. Because Ollama runs offline, responses are fast even on consumer hardware.

Pro Tips

• Use the “Continue” button to have the AI extend its last reply.
• Regenerate to get alternative responses.
• Edit the card mid-chat to adjust her personality on the fly.

---

Enhancing Your Workflow with MiniTavern

While SillyTavern and Ollama are a powerful combination, managing character cards across devices can be a hassle. That’s where the MiniTavern ecosystem shines.

• MiniTavern Mobile Apps (iOS/Android) let you carry your character cards on the go. Import cards from your phone and chat with them using your local Ollama instance.
• Web Tavern provides a browser-based interface for quick access without installing anything.
• Chrome Extension integrates character cards directly into your browsing experience, letting you summon Elena or any other character in any text field.
• Character Card Market is the largest curated collection of high-quality character cards. You’ll find expert-crafted cards like Elena, complete with rich backstories and tested dialogue formats. New cards are added weekly.

By combining these tools, you can build a seamless workflow: download cards from the market, import them into SillyTavern, and chat locally with Ollama—then sync your favorites to your phone for on-the-go roleplay.

---

Troubleshooting Common Issues

• Ollama not connecting – Ensure both apps are running and the API URL is correct. Check your firewall.
• Model doesn’t follow the card – Lower the temperature and increase the context size. Some models need a system prompt like “You are Elena the Enchantress. Stay true to her description.”
• Slow responses – Use a smaller model or reduce the context size. Close other GPU-intensive apps.
• Card won’t import – Verify the file is a valid character card format. Try using the MiniTavern Chrome extension to export cards in the correct format.

---

Conclusion

Running SillyTavern character cards with Ollama in 2026 gives you complete control over your AI roleplay experience. You get privacy, zero costs, and the ability to craft rich, consistent characters like Elena the Enchantress. By following this guide, you’ve set up a local LLM pipeline that’s both powerful and easy to maintain.

To take your character card collection further, explore the MiniTavern ecosystem. Download the mobile apps to chat anywhere, use the Web Tavern for instant access, install the Chrome extension for seamless integration, and browse the Character Card Market for professionally designed cards. Whether you’re a writer, gamer, or just curious about AI storytelling, this setup will unlock hours of creative conversation.

Happy roleplaying