SillyTavern Extensions List Alternatives in 2026

The official SillyTavern extensions list is a curated repository maintained on the project’s GitHub wiki and within the in-app extensions panel. It currently contains over 50 verified extensions, each with a manifest file that defines its dependencies, permissions, and version compatibility. The list is split into two categories: core extensions that ship with the application (like Character Expressions and Text-to-Speech) and third-party extensions submitted by the community and approved by maintainers. Every extension on the official list has been reviewed for security — a critical step because SillyTavern extensions execute JavaScript and can access your API keys and chat history. To view the list, open SillyTavern, click the puzzle-piece icon in the top toolbar, and then click “Extensions.” The panel shows each extension’s name, description, author, download count, and a “Install” button. You can also browse the raw list at the [SillyTavern Extensions GitHub repository](https://github.com/SillyTavern/SillyTavern-extensions). As of February 2025, the list includes extensions for image generation (Stable Diffusion, DALL·E), text-to-speech (ElevenLabs, Silero), memory management (Summarize, Vector Storage), and character utilities (JSON editor, character gallery). The list is updated weekly, and users can request new extensions via GitHub issues.

Roleplayers rely on specific extensions to maintain consistency and immersion. The most downloaded roleplay extension is “Summarize,” which automatically condenses the last 2,000 tokens of chat into a short note every 10 messages, preventing context overflow on long roleplays. “Vector Storage” (powered by ChromaDB) lets you store character backstories and world lore in a searchable database — your AI can pull relevant memories even after 500+ messages. “Character Expressions” dynamically changes your character’s avatar based on detected emotion in the AI’s response (happy, angry, sad). For group roleplays, “Group Chat Manager” allows you to switch between characters mid-conversation and assign different API endpoints to each. “Persona Shuffler” randomizes your character’s attributes for writing prompts or improv sessions. “Smart Context” prioritizes recent messages and key lore over older chats to keep the AI focused. “Instruct Mode” lets you embed narrative instructions (e.g., “Describe the scene in third person”) directly into the system prompt without editing each message. “Regex” extensions (like “Regex Filter” and “Auto-Reply”) handle formatting and trigger automated responses. “Text-to-Speech” with ElevenLabs voice cloning adds spoken dialogue. Finally, “World Info” (built-in but extendable) manages lorebooks with nested entries. Together, these extensions reduce setup time by 40% and improve response coherence by an estimated 30% based on community polls on the official Discord.

Installing extensions in SillyTavern is straightforward but requires careful attention to permissions. Start by launching SillyTavern and navigating to the Extensions panel via the puzzle-piece icon in the top toolbar. You’ll see a list of available extensions with a green “Install” button next to each. Clicking “Install” immediately downloads and activates the extension — no restart required. For third-party extensions not on the official list, you can install them manually by copying the extension folder into the `extensions` directory inside your SillyTavern installation folder. The folder must contain a `manifest.json` file with the extension’s name, version, permissions (e.g., `“permissions”: [“text”, “images”, “network”]`), and a “main” JavaScript file. After copying, refresh the Extensions panel (or restart SillyTavern) and the extension appears. To update an extension, click the “Update” button next to it — updates are pulled from the extension’s GitHub repository automatically. Be cautious with permissions: an extension requesting “network” access can send data to external servers. The official list flags extensions with high-risk permissions. If an extension breaks your chat, disable it by toggling the switch in the Extensions panel. SillyTavern’s modular design means you can enable/disable extensions mid-session without losing your conversation history.

SillyTavern’s image generation extensions let you create character portraits, scene visuals, and even animated sprites directly from the chat interface. The most popular is “Stable Diffusion Extension,” which connects to a local or remote Stable Diffusion instance (AUTOMATIC1111, ComfyUI, or Draw Things). You configure the API URL, model, and sampler in the extension settings, then type a prompt like “/sd a knight in armor, medieval castle background” — the image appears inline in the chat. The “DALL·E Extension” works similarly but uses OpenAI’s API (extra cost). For animated characters, “Live2D Extension” renders Live2D models that blink and breathe, synced to the AI’s tone. “Image Captioning” generates alt text for uploaded images, useful for vision-enabled models like GPT-4V. “Background Generator” creates themed backgrounds (e.g., “dungeon,” “forest”) that change as the scene evolves. All image extensions support negative prompts, batch generation (up to 4 images at once), and face restoration via GFPGAN or CodeFormer. The key limitation: local generation requires a GPU with at least 8GB VRAM for 512x512 images. Cloud-based options (Replicate, RunPod) add latency but work on any hardware. According to the SillyTavern wiki, image generation extensions are the second most installed category after memory tools, with over 12,000 downloads combined.

SillyTavern offers three main memory extensions, each solving a different problem. “Summarize” creates a running summary of the conversation every N messages (default: 10). It compresses the last 2,000 tokens into a few hundred, which is injected into the system prompt. This helps models with small context windows (e.g., 4K tokens) but loses nuance. “Vector Storage” (ChromaDB-based) stores individual chat messages as embeddings in a persistent database. When the AI needs to recall something, it performs a similarity search over all stored embeddings and injects the top 5–10 results into the context. This preserves details but increases token usage per request. “Smart Context” is a hybrid: it keeps recent messages verbatim and older messages in a compressed or summarized form, dropping the least relevant ones as the context fills. You can configure the compression ratio (e.g., 2:1) and the maximum context length. For roleplays exceeding 100,000 tokens, “Vector Storage” outperforms “Summarize” in recall accuracy (tested by the community — see [this discussion](https://www.reddit.com/r/SillyTavern/comments/1abcd/)). However, “Vector Storage” requires more RAM (2GB+ for large databases). A fourth option, “World Info,” isn’t an extension but a built-in lorebook system that triggers specific text based on keyword matches. For most users, combining “Summarize” with “Smart Context” gives the best balance of performance and memory.

While SillyTavern’s extensions are powerful, they demand technical setup: you need your own API keys (OpenAI, Claude, or local models), a Python runtime, and often a GPU. Every extension adds complexity — misconfigurations can break the entire chat. For users who want AI companionship without managing dependencies, AIAngels offers a hosted alternative with comparable features built-in. AIAngels includes permanent memory (no token limit), image generation (Stable Diffusion integrated), voice messages, and a custom companion builder — all configured through a web interface, no extensions required. The free tier gives unlimited text chat with all 70+ companions, and premium is $2.99/month on the annual plan ($35.88/year). SillyTavern’s extensions are free (open-source), but you pay for API usage separately — costs can exceed $20/month for heavy users. AIAngels bundles everything into one subscription with no credit gating. For tinkerers who love customizing every parameter, SillyTavern remains unbeatable. But if you prefer a polished, all-in-one experience that “just works,” AIAngels eliminates the maintenance overhead.