Self-host SillyTavern
1.19.0LLM frontend for power users with unified API interface
- Stars
- 35 646
- CPU
- Any CPU capable of running Node.js 20
- RAM
- Minimal for the frontend; 6GB+ VRAM recommended for local LLM inference
- Storage
- Minimal for application data
- Port
- 8000 (HTTP server)
ghcr.io/sillytavern/sillytavernWhat is SillyTavern?
SillyTavern is a self-hosted web application that serves as a unified frontend for interacting with large language models. It connects to numerous LLM APIs including OpenAI, Anthropic Claude, OpenRouter, KoboldAI, NovelAI, and local backends like Ollama and llama.cpp. Users create and manage character cards, maintain chat histories, and use features like Visual Novel Mode, WorldInfo (lorebooks), automatic translation, and text-to-speech. The UI is mobile-friendly and extensible through third-party extensions and server plugins. It runs on Node.js 20+ and can be installed via Docker, native scripts on Windows/Linux/macOS, or Termux on Android. The project is licensed under AGPL-3.0 and developed by the SillyTavern organization.
Should you self-host SillyTavern?
SillyTavern scores 83/100 on selfhost.directory (how scoring works) and ranks #3 of 68 in AI Chat. It has 35 646 GitHub stars, +1 315 in the last 30 days.
It shipped 11 stable releases in the past year. The latest, 1.19.0, came out on 14 September 2026.
Setup is moderate: expect some configuration beyond a single command. An official Docker image is available. It is AGPL-3.0-licensed and written mostly in JavaScript.
Its closest alternative on selfhost.directory is LibreChat (85/100, 47 150 stars). It scores higher overall, so compare the two before you pick. See SillyTavern vs LibreChat side by side.
Features
- Auto-Translation
- Character Management
- Chat Backups
- Customizable UI
- Extensions & Plugins
- Image Generation Integration
- IP Whitelisting
- Multi-LLM API Support
- Multi-User Mode
- SSO Integration
- Text-to-Speech (TTS)
- Visual Novel Mode
- WorldInfo / Lorebooks
SillyTavern README
How to install SillyTavern
Install SillyTavern with Docker
SILLYTAVERN_VERSION, PUBLIC_PORT, CONFIG_PATH, DATA_PATH, PLUGINS_PATH, EXTENSIONS_PATHdocker compose up
Uses prebuilt image from GitHub Container Registry
Install SillyTavern with Linux
git clone https://github.com/SillyTavern/SillyTavern -b release cd SillyTavern ./start.sh
Requires Node.js 20+ and Git
Install SillyTavern with Windows
git clone https://github.com/SillyTavern/SillyTavern -b release cd SillyTavern double-click Start.bat
Requires Node.js 20+ and Git for Windows. Do not install in Windows controlled folders.
Install SillyTavern with macOS
git clone https://github.com/SillyTavern/SillyTavern -b release cd SillyTavern ./start.sh
Requires Node.js 20+ and Git
SillyTavern releases & version history
AI summary Added support for new AI models
SillyTavern 1.19.0
Backends
- Anthropic: Added support for Claude Fable 5 and 5.1, Claude Opus 4.8 and 5, Claude Sonnet 5.
- OpenAI: Added support for GPT-5.6 models, GPT-6 Astra.
- Other providers: Added support for Gemini 3.5/3.6/3.7 Flash models, GLM-5.2, and DeepSeek V4 Flash Vision Exp.
- Google AI Studio: Model list now fetches all pages.
- Fireworks AI: Updated model fetching; added reasoning effort, reasoning display, and session affinity for prompt caching.
- OpenRouter: Added optional logprobs support and refreshed provider list.
- NanoGPT: Refreshed provider list.
- Pollinations: Added keyed and keyless endpoint selection; updated TTS model aliases.
- DeepSeek: Added low reasoning effort.
- Generic Text Completion: Model list is now shown in a select.
- NovelAI: Preserve zero temperature for Math1; updated account information endpoint.
- Updated Gemini TTS to Gemini 3.1 Flash TTS Preview, added GLM-5.2 to Z.AI, removed MiniMax TTS GroupId requirement, and updated BFL API endpoints.
- Custom OpenAI-compatible sources: Added macro substitution in custom headers and request bodies; improved remote tokenizer validation.
Improvements
- World Info: "Apply Current Sorting" now supports ascending and descending order, configurable start and step values, and live validation.
- World Info: Renaming lorebooks now updates chat, character, and persona links.
- Chat Completion: Added an expand editor button to quick prompts.
/addswipeno longer reloads the entire chat.- Chat files with invalid headers are no longer silently overwritten, and chats with an invalid final line remain accessible with a degraded preview.
- New chat branches and checkpoints now receive fresh integrity identifiers.
- Non-ASCII chat names now use distinct backup keys and throttles.
- Deleting messages now also removes associated tool-call messages.
- Improved performance when sorting World Info entries, listing plugins, sorting images by date, and handling autocomplete on mobile.
Macros
- Variable macros can now access array elements and object properties.
- Added
{{defaultExpression}},{{lastExpression}}, and{{availableExpressions}}macros for Character Expressions. - Fixed scoped comment macros.
- Fixed pipe characters in macro arguments while pipe/filter tokenizing remains disabled.
- Fixed trailing whitespace in variable macros.
STscript
/expression-listcan filter custom expressions and supports additional return formats.- Fixed inflated
/tokensresults for OpenAI tokenizers. - Fixed event type emitted by
/note-position.
Extensions
- Added
MessageFormatter, allowing extensions to transform message content before regex processing, after regex processing, or after Markdown conversion. - Character Expressions: Improved handling of custom expressions and fallback values.
- ComfyUI: History lookup now uses only the current prompt and ignores non-image outputs.
- Fixed multimodal caption requests for Text Generation WebUI.
- Extension discovery now ignores folders without a manifest.
Server & Security
- Blocked
localhostand.localhostaliases from bypassing private-address checks in/api/search/visit. - Added rate limiting to account reset requests.
- Tightened npm package source restrictions.
- Fixed CORS proxy responses retaining compression headers after decompression.
- Fixed boolean parsing in environment variable migrations.
Bug Fixes
- Fixed connection profile switching leaving the previous Chat Completion source active.
- Fixed duplicate tool-call IDs when providers repeat tool metadata while streaming.
- Provider error responses returned with HTTP 200 now show the provider error instead of the HTTP status text.
- Fixed Quick Reply overwrite cancellation and replacement dialog label.
- Fixed whitespace-only reasoning blocks leaving visible controls.
- Fixed group avatars ignoring the "Never resize avatars" setting.
- Fixed tag filters remaining visible after their last assigned entity is removed.
- Fixed Character Expressions errors with missing default-expression labels.
- Fixed character renaming when an optional source directory does not exist.
- Fixed Chromium password managers detecting proxy access-key fields as passwords.
- Fixed Chub URL parsing and stale autocomplete observers.
- Fixed crashes when a chat is deleted during search or recent-chat scanning.
- Fixed BYAF imports reading unrelated bytes from pooled buffers.
Community updates
- Refactor getUniqueName utility for allowBaseName option by @Wolfsblvt in #5602
- fix /tokens command returning inflated count for OpenAI API by @Reithan in #5601
- Add expand editor button to CC quick prompts by @CasualAutopsy in #5607
- Fix stale chat bound lore name after rename by @unnaturaldeveloper in #5610
- Fix emitted event type in note-position command by @Cohee1207 in #5624
- feat(openai): resolve macros in Custom additional body, exclude body, and headers by @manthanawgan in #5627
- Fix mobile lag from inactive autocomplete resize handlers by @Tantoofaaz777 in #5635
- Make parseChubURL clean URLs before relying on length by @Rojikku in #5604
- Feat/update persona lore by @Cohee1207 in #5637
- Feat/New features for character expressions by @leandrojofre in #5574
- Add message indicating that the 'Request mode reasoning' checkbox controls whether reasoning is generated at the source, for Deepseek API by @Feldherren in #5633
- fix mislabelled popup for "Replace existing Quick Reply set" by @underscorex86 in #5662
- chore: Sync NanoGPT providers: Tensorix, and Xiaomi by @DeathStalker471 in #5646
- Update outdated readme-zh_tw.md by @40436michael in #5639
- Update outdated readme-zh_cn.md and readme-ja_jp.md by @realyxl in #5659
- fix(autocomplete): guard MutationObserver against detached textarea by @mekineer-com in #5660
- Fix: Disable Incomplete Pipe/Filter Tokenizing to Prevent Breaking Macros with Pipe Characters in Arguments by @Wolfsblvt in #5644
- Lorebook "Apply Current Sorting" — Order control, ascending mode, step size, and live validation feedback by @Wolfsblvt in #5653
- Fix: Scoped comment macro
{{//}}...{{///}}not removing its content by @Wolfsblvt in #5643 - Add
MessageFormatter: a structured hook system for extensions to transform message content pre-DOM by @Wolfsblvt in #5652 - Pollinations: Add API endpoint selection by @CloudCompile in #5625
- Enforce stricter npm source restrictions in .npmrc by @Cohee1207 in #5663
- Add rate limit to account reset by @Cohee1207 in #5603
- Macros: Access array and object elements (#4383) by @blueberrys in #5649
- Improve remote tokenizer encode validation by @Cohee1207 in #5665
- Temporarily comment out allow-remote setting in .npmrc due to npm bug by @Cohee1207 in #5668
- Add gemini-3.5-flash to Google AI Studio and Vertex model lists by @Tosd0 in #5675
- change novelai user information to be called from image.novelai.net by @underscorex86 in #5814
- Add glm-5.2 to contextMap by @Samueras in #5770
- Update TTS model from Gemini 2.5 to 3.1 in TTS Extension by @Samueras in #5778
- Update TTS model from Gemini 2.5 to 3.1 in TTS Extension (#5778) by @Wolfsblvt in #5824
- Add claude-fable-5 support to Claude and OpenAI-compatible request builders by @TheLonelyDevil9 in #5757
- Openrouter: Enable logprobs if requested by user by @lukasstraub2 in #5804
- Optimize World Info entry sorting to avoid O(n^2) lookups by @ransxd in #5809
- Fix translation for 'Persona Description' in French by @maxime-fleury in #5830
- Add Claude Opus 4.8, GPT-5.6 models, and migrate Gemini 3.1 Flash Lite and Gemini image models to stable by @Jasonnor in #5845
- Fix: ignore local third-party extensions during core linting by @Wolfsblvt in #5847
- Fix: remove trailing whitespace from variable macros by @Wolfsblvt in #5848
- Fix whitespace-only reasoning block controls by @DragonFSKY in #5636
- Sync OpenRouter providers list by @cloak1505 in #5686
- chore: Sync NanoGPT providers: UOMI and Wafer by @DeathStalker471 in #5707
- perf(util): read each file's mtime once when getImages sorts by date by @tgies in #5689
- perf(plugins): list plugin dirs without per-entry statSync by @tgies in #5690
- fix: fetch only current prompt from ComfyUI history endpoint by @frankschubert in #5790
- Fail chat integrity check on unparseable headers instead of silently overwriting by @Wolfsblvt in #5928
- fix: reject localhost and .localhost domains in /api/search/visit to prevent SSRF bypass by @kings9527 in #5749
- Fix connection profile switch leaving the previous chat completion source active by @Wolfsblvt in #5930
- Update BFL generation and polling endpoints to api.bfl.ai by @AndSi-327 in #5681
- DeepSeek: Add/enable low reasoning effort by @0x4kgi in #5924
- Fix crash when a chat is deleted during a search/recent scan by @Max-Gomez in #5871
- Don't do a full chat reload from /addswipe by @alpha-omicron in #5903
- Verify GPT-5.6 support and refresh OpenAI model list by @Wolfsblvt in #5932
- Add Claude Opus 5 / Sonnet 5 and Gemini 3.7 Flash / 3.6 Flash / 3.5 Flash-Lite by @Jasonnor in #5883
- fix: skip non-image ComfyUI outputs by @CBOSSX in #5910
- fix: use exact ArrayBuffer slices for BYAF imports by @Moli13337 in #5917
- fix: remove GroupId requirement from MiniMax TTS by @guozi-lab in #5920
- fix: show full Generic textgen model list in a select by @Trahloc in #5897
- Skip extension folders without a manifest during discovery by @Wolfsblvt in #5934
- fix(tags): drop filters for tags with no assigned entities by @mtibke in #5911
- Fireworks: Update model fetching because /v1/models is broken by @Christoph-D in #5831
- fix: correct malformed multimodal caption request for oobabooga by @bbastex in #5915
- Prevent proxy access keys from triggering password-manager prompts in Chromium-based browsers by @CodeBoy2006 in #5913
- fix(expressions): guard against null/undefined default-expression labels by @Lockyer228 in #5938
- fix: never resize avatars option does not work for group chat avatar by @m-petra-fn in #5936
- Mint a fresh integrity slug for new branches and checkpoints by @Wolfsblvt in #5943
- Give non-ASCII chat names distinct backup keys and throttles by @Wolfsblvt in #5944
- Use provider error fields instead of statusText for HTTP 200 errors by @Wolfsblvt in #5945
- Fireworks AI: Set session affinity header for improved caching by @Christoph-D in #5826
- Fireworks: Request and process reasoning by @Christoph-D in #5947
- Fix duplicate tool call IDs if provider repeats id, type, name by @Christoph-D in #5948
- fix: parse boolean values in env migrations by @primorLee in #5950
- fix(chat): delete associated tool calls with messages by @LazyPeterGriffin in #5949
- Update TavernAI link in README by @m98bw in #5987
- fix: preserve zero NovelAI Math1 temperature by @tandede in #5965
- Return a degraded preview instead of dropping a chat whose last line is unparseable by @shoemoney in #5969
- Google AI Studio: Fetch all model pages by @saitewasreset in #5954
- Fix Quick Reply overwrite cancellation by @QiuLsG in #6005
- fix: resolve Pollinations TTS model aliases by @voodoohop in #6010
- Fix: Avoid cpSync crash when renaming characters by @chansey97 in #6012
- Fix CORS proxy forwarding raw zstd-compressed bytes to the client by @X00LA in #6007
- Add image embed for DeepSeek V4 Flash Vision Exp (#5964) by @fishBone000 in #5996
- Add Claude Fable 5.1 and GPT-6 Astra support by @Wolfsblvt in #6015
- Update release version number by @Wolfsblvt in #6035
- 1.19.0 by @Wolfsblvt in #6036
New Contributors
- @manthanawgan made their first contribution in #5627
- @Tantoofaaz777 made their first contribution in #5635
- @Rojikku made their first contribution in #5604
- @Feldherren made their first contribution in #5633
- @40436michael made their first contribution in #5639
- @realyxl made their first contribution in #5659
- @mekineer-com made their first contribution in #5660
- @CloudCompile made their first contribution in #5625
- @blueberrys made their first contribution in #5649
- @TheLonelyDevil9 made their first contribution in #5757
- @lukasstraub2 made their first contribution in #5804
- @ransxd made their first contribution in #5809
- @maxime-fleury made their first contribution in #5830
- @DragonFSKY made their first contribution in #5636
- @frankschubert made their first contribution in #5790
- @kings9527 made their first contribution in #5749
- @AndSi-327 made their first contribution in #5681
- @0x4kgi made their first contribution in #5924
- @Max-Gomez made their first contribution in #5871
- @alpha-omicron made their first contribution in #5903
- @CBOSSX made their first contribution in #5910
- @Moli13337 made their first contribution in #5917
- @guozi-lab made their first contribution in #5920
- @mtibke made their first contribution in #5911
- @bbastex made their first contribution in #5915
- @CodeBoy2006 made their first contribution in #5913
- @Lockyer228 made their first contribution in #5938
- @primorLee made their first contribution in #5950
- @LazyPeterGriffin made their first contribution in #5949
- @m98bw made their first contribution in #5987
- @tandede made their first contribution in #5965
- @shoemoney made their first contribution in #5969
- @saitewasreset made their first contribution in #5954
- @QiuLsG made their first contribution in #6005
- @voodoohop made their first contribution in #6010
- @chansey97 made their first contribution in #6012
- @fishBone000 made their first contribution in #5996
Full Changelog: 1.18.0...1.19.0
SillyTavern 1.18.0
Important news
Read the maintainers statement regarding a recent security incident involving the "Bot Browser" third-party extension and learn how to stay safe: #5592
Backends
- Added Cloudflare Workers AI and MiniMax as Chat Completion sources.
- KoboldCpp: Grammar state will be preserved when using a "Continue" option.
- KoboldCpp: Added forwarding of reasoning effort when running as a Custom Chat Completion source.
- Tool Calling: Added a configurable tool calling recursion limit; enabled interleaved thinking for Custom sources.
- Text Completion: Impersonation requests use a "Last User Message" prefix at the end of the prompt (if configured).
- Text Generation WebUI: Added Adaptive-P controls.
- NanoGPT: Added provider selection and model sorting.
- Added ability to view remaining balance for OpenRouter and NanoGPT.
- Enhanced support for new models: DeepSeek v4, GPT 5.4 and 5.5, Gemma 4, GLM-5V-Turbo, Claude Opus 4.7.
Server & Security
- Removed post-install script, config migration is now handled by the app or a dedicated
npm run initcommand. - Added npm configuration to prevent execution of package scripts during installation.
- Moved HTTP error pages and
user.cssfile from/publicto/datato support immutable setups. - Disabled HTTP keep-alive by default to restore old Node 18 behavior, can be enabled with config.
- Added rate limiting to the basic authentication flow to mitigate brute-force attacks.
- Added configuration options to choose which headers can be used for forwarded IP detection to prevent spoofing.
- Added a private address whitelist to prevent SSRF attacks. See the documentation on how to enable and configure: Private Address Whitelist.
- Added an IP whitelist for SSO trusted proxies to prevent authentication bypass.
- Added invalidation of session cookies on password change to prevent session hijacking.
- Increased the length of password reset code to 6 characters to guard against brute-force attacks.
- Implemented PKCE challenge in OpenRouter OAuth flow for more secure key exchange.
UI/UX
- Improved swipe picker: mobile requires a long press on swipe counter to open; added buttons to expand or copy the swipe text.
- "Click to Edit" mode now also applied to reasoning blocks.
- Welcome Screen: Number of recent chats can be configured.
- Streamed requests now can show an error message in the console if the request fails.
STscript
- Added commands for persona management:
/persona-create,/persona-update,/persona-delete,/persona-duplicate, and/persona-get. - Added a command to force update the Prompt Manager's prompt list:
/pm-render. - Added a command to get the state of the regex script:
/regex-state. - Added a command to set fallback expression:
/expression-fallback. - Added a command to generate a streamed response with a connection profile:
/profile-genstream.
Extensions
- Assets list now groups extensions by "Official" or "Community" categories.
- Added an additional confirmation prompt when installing third-party extensions (can be disabled).
- Supported extensions can use a secret-id from connection profiles when making an LLM request.
- Extensions list now shows the extension's author name resolved from the git remote URL.
- Vector Storage: Added Workers AI source; added a toggle to keep vectors for hidden messages; added retry logic to summary generation.
- Image Generation: Added Workers AI source; generation can now be cancelled by pressing a button in the status toast.
- Image Captioning: Added support for macros in the caption prompt.
- TTS: "Skip code blocks" no longer ignores lines that start with 4 spaces (legacy code block syntax); "disabled" voice now shows a toast only once per character.
Bug Fixes
- Fixed text edit flow in Firefox on mobile.
- Fixed welcome screen chat pins not updating on chat renaming.
- Fixed character list filters being stuck on app initialization.
- Fixed application of instruct formatting to
/genrawrequests. - Fixed model routing to sd.cpp API in Image Generation logic.
- Fixed validation of image URLs generated with Z.AI API.
- Fixed vectors deletion for KoboldCpp when a message is deleted.
- Fixed "Show More Messages" button triggering edit in "Click to Edit" mode.
- Fixed max height of select-multiple elements in mobile layout.
- Fixed server crash on empty messages when applying cache control parameters.
Community updates
- Fix npm audit in /tests by @Cohee1207 in #5370
- CI: Update node to 24, run ci without scripts by @Cohee1207 in #5376
- Backmerge release into staging by @Cohee1207 in #5377
- feat(secrets): update readSecret function to accept optional secret ID by @dylenyedc in #5356
- fix: return Error objects from invokeFunctionTool and create error invocations by @xz-dev in #5351
- Add opt-in toggle to keep hidden messages in chat vector index by @Copilot in #5378
- feat: swipe picker expand/collapse, copy button, and left-align text by @Copilot in #5380
- Fix /genraw user instruct format not applied and unwanted system newline by @claude in #5372
- fix: require long press to open swipe picker on phones by @Cohee1207 in #5382
- Save chat before emitting event for user message by @Cohee1207 in #5389
- Use custom init script instead of postinstall by @Cohee1207 in #5384
- glm-5v-turbo by @Cohee1207 in #5393
- Add VSCode settings for TypeScript SDK configuration by @Cohee1207 in #5394
- Add Gemma 4 instruct, context, and reasoning presets by @LumiWasTaken in #5403
- Immutable public and global content management by @Cohee1207 in #5390
- Skip TTS for voices explicitly set to disabled (fixes #4970) by @tgies in #5367
- Add unit tests for prompt-converters.js by @tgies in #5364
- Add unit tests for TavernCardValidator by @tgies in #5366
- Add tests for Cache, MemoryLimitedMap, and other util.js coverage gaps by @tgies in #5365
- Add IP whitelist for SSO authentication headers by @Cohee1207 in #5404
- fix: improve streaming error propagation and forwarded response logging by @dz114879 in #5317
- feat: add Cloudflare Workers AI provider by @tgies in #5385
- fix npm audit in src/electron by @Cohee1207 in #5405
- Add Entry Preview to World Info Deletion Confirmation Dialog by @Wolfsblvt in #5423
- refactor(tts): remove redundant 4-space codeblock filter by @bumprat in #5406
- feat: add Workers AI text embeddings and multimodal captioning by @tgies in #5414
- Fix: Remove
COMFY_RUNPODfrom secrets input map — invalid element reference by @Wolfsblvt in #5432 - Fix: Suppress false "Unsupported language" console warnings for English locale variants by @Wolfsblvt in #5431
- Fix: Skip Pseudo-Elements When Generating Dynamic Focus Styles by @Wolfsblvt in #5430
- Add Arguments Logging to Deprecated MacrosParser Method Warnings for Better Debugging Context by @Wolfsblvt in #5429
- Fix stable-diffusion.cpp model routing and URL path handling by @V13Axel in #5427
- Add
/regex-stateslash command to check regex script status by @Wolfsblvt in #5428 - Migrate built-in extensions to use manifest-based activate hooks by @Wolfsblvt in #5435
- Refactor repetitive vectorization model loading into a generic data-driven function by @Copilot in #5425
- Add configurable welcome screen recent chats, fix pin loss on rename, and add popup number input validation by @Copilot in #5436
- Gemma 4 reasoning: drop newline there by @Cohee1207 in #5433
- Chore: Add persona lifecycle events (
PERSONA_CREATED,PERSONA_UPDATED,PERSONA_RENAMED,PERSONA_DELETED) by @Wolfsblvt in #5443 - chore: Add
getExtensionManifest()to the extension context API by @Wolfsblvt in #5442 - Add
disabledproperty support to all popup custom input types by @Wolfsblvt in #5441 - Fix: Missing
signatureandtoolSignaturesfields inChatCompletionServicestreaming state by @Wolfsblvt in #5439 - Allow grammar to keep place during continue by @Reithan in #5444
- Add
cleanextension lifecycle hook for optional data cleanup by @Wolfsblvt in #5449 - Emit
PERSONA_CREATEDevent on persona duplication by @Wolfsblvt in #5448 - fix: missing closing bracket in jQuery selector in setInContextMessages() by @Copilot in #5451
- Add Streaming Display Utility and New Generation Slash Commands (
/genstream,/reasoning-format) by @Wolfsblvt in #5438 - Integrate Cloudflare Workers AI text-to-image into SD extension by @Copilot in #5434
- Use 'Last User Prefix' for Last Prompt Line When Sending TC Impersonation Requests by @CasualAutopsy in #5456
- Enable interleaved tool reasoning for custom OpenAI-compat endpoints by @Reithan in #5445
- Improve OpenRouter model lists in extensions by @Cohee1207 in #5459
- Enhanced
/persona-syncCommand with Range, Name Filter, and Quiet Mode by @Wolfsblvt in #5460 - implemented emit events for itemized-prompts.js by @Enerccio in #5461
- Fix incorrect translation of "New" button in World Info module by @awaae001 in #5463
- Increase Top K max value from 200 to 500 for text generation WebU by @Wolfsblvt in #5474
- add claude-opus-4-7 by @ashishch432 in #5465
- Fix group chat reasoning "mind reading" — only include reasoning from the currently generating character by @Wolfsblvt in #5473
- fix: limit max-height of select[multiple] on mobile to prevent Chrome 145+ full-screen rendering by @octo-patch in #5467
- fix: correct typo in cacheScores debug log and add JSDoc to filterDataByState by @octo-patch in #5468
- Bulk extension field updates via merge-attributes with UNSET_VALUE sentinel by @Wolfsblvt in #5471
- Add 'pm-render' command to refresh prompt manager content by @Cohee1207 in #5483
- fix: enhance URL validation for Z.AI image generation by @Cohee1207 in #5482
- Refactor: Replace SD image generation indicator with ActionLoader system by @Wolfsblvt in #5472
- Fix: Prevent "Show More Messages" Button from Triggering Message Edit Mode by @Wolfsblvt in #5486
- Add Click-to-Edit Support for Reasoning Blocks by @Wolfsblvt in #5487
- Add Persona CRUD Slash Commands with Shared Avatar Utilities by @Wolfsblvt in #5466
- Enhance Vectorize All process with error handling, retries and minor improvements by @TanJeeSchuan in #5479
- Add Slug Parameter to Action Loader for Programmatic Identification by @Wolfsblvt in #5490
- feat: add Adaptive-P controls for oobabooga by @yuaiccc in #5502
- Sync OpenRouter providers list by @cloak1505 in #5503
- feat: add getWorldInfoNames() to getContext() for WorldInfo enumeration. by @Therosin in #5505
- Fix KoboldCpp vector hash deletion by @yuaiccc in #5508
- Display OpenRouter credit balance in UI by @Stagnating in #5513
- feat: add MiniMax as a chat completion provider by @octo-patch in #5452
- feat: add DeepSeek V4 model support with thinking mode and reasoning effort by @Dclef in #5522
- fix: add config toggle to disable HTTP keepAlive (Node 18 behavior) by @DeathStalker471 in #5519
- fix: Don't apply layout hack in Firefox Mobile while editing text by @Christoph-D in #5531
- add macro support for image caption extension prompt (multimodal) by @f-eg in #5527
- Add Tool Call Recurse Limit slider to Chat Completion settings by @Copilot in #5518
- OpenAI: Add gpt-5.5, gpt-5.4-mini/nano, gpt-image-2 by @Cohee1207 in #5529
- feat: add gemma 4 for AI studio by @Cohee1207 in #5493
- feat: pass koboldcpp reasoning effort by @Cohee1207 in #5491
- fix: remove legacy filter state restoration by @Cohee1207 in #5535
- Implement S256 challenge in OpenRouter OAuth flow by @Cohee1207 in #5501
- Chat Completion: common model sorting and grouping settings, improved NanoGPT model list by @DeathStalker471 in #5536
- feat: Add NanoGPT credit stats UI by @DeathStalker471 in #5537
- Fix missing filename sanitization on V2 JSON character import + harden getPngName as safety nee by @Wolfsblvt in #5538
- Implement private IP range request host validator by @Cohee1207 in #5497
- Fix: Prevent crash in cachingAtDepthForOpenRouterClaude on empty content from trailing tool calls by @crsp6447 in #5541
- Better SVG for Kobold/KoboldCpp by @bubfix in #5543
- fix: improve sanitation of toasts that bypass HTML escaping by @Cohee1207 in #5540
- fix: update backup archive to ignore migration secrets files by @Cohee1207 in #5548
- Backmerge release into staging by @Cohee1207 in #5549
- Minor adjustments to the SVG icons. by @bubfix in #5546
- fix: (AI Studio) temperature -> maxTemperature by @Cohee1207 in #5554
- Extension management improvements by @Cohee1207 in #5552
- Fix json schema use for openAI compat CUSTOM endpoints in several use paths by @Reithan in #5561
- Feat - Add expression-set-fallback slash command by @leandrojofre in #5551
- feat: add nanogpt provider selection by @DeathStalker471 in #5544
- Add rate limit to basic auth middleware by @Cohee1207 in #5504
- Update/Turn expression-set-fallback into expression-fallback by @leandrojofre in #5564
- Add account version tags to cookies by @Cohee1207 in #5563
- Extension clone improvements (part 2) by @Cohee1207 in #5571
- fix: npm audit package dependencies by @Cohee1207 in #5572
- Add option to return malformed JSON string from extractJsonFromData by @Cohee1207 in #5578
- Refactor extension management and assets download menu by @Cohee1207 in #5583
- Tabby: Prevent unwanted BOS token being added to all tokenizations. by @Ph0rk0z in #5589
- Update release version number by @Cohee1207 in #5590
- Staging by @Cohee1207 in #5591
New Contributors
- @dylenyedc made their first contribution in #5356
- @dz114879 made their first contribution in #5317
- @bumprat made their first contribution in #5406
- @V13Axel made their first contribution in #5427
- @Enerccio made their first contribution in #5461
- @ashishch432 made their first contribution in #5465
- @octo-patch made their first contribution in #5467
- @TanJeeSchuan made their first contribution in #5479
- @yuaiccc made their first contribution in #5502
- @Stagnating made their first contribution in #5513
- @Dclef made their first contribution in #5522
- @DeathStalker471 made their first contribution in #5519
- @f-eg made their first contribution in #5527
- @crsp6447 made their first contribution in #5541
- @bubfix made their first contribution in #5543
Full Changelog: 1.17.0...1.18.0
SillyTavern 1.17.0
Note: Requires Node.js 20 or higher.
Need to update Node? See: https://docs.sillytavern.app/installation/updating/node/
Backends
- Claude: Added adaptive thinking support mapped from Reasoning Effort (opt-in via config)
- OpenRouter: Added provider filtering based on selected model
- OpenRouter: Deactivated reasoning toggle with minimal reasoning effort now disables thinking (model-dependent)
- OpenRouter: Added interleaved reasoning forwarding for tool-call continuations
- SiliconFlow: Added API endpoint selection (Global / China)
- xAI: Removed deprecated web search toggle
- Synchronized model lists (GPT, Claude, GLM, Gemini, Grok Imagine)
Improvements
- Swipe Picker: New interface to browse, jump, branch, and delete swipes
- Backgrounds: Added virtual folders with grid view, breadcrumbs, and thumbnails
- UI: New splash screen design during app initialization
- World Info: Renaming can optionally relink lorebooks across all characters
- World Info: Unified chat, character and persona link buttons behavior (click to load, shift-click or long press to open a menu)
- Tags: Added automatic cleanup of orphaned folder tags
- Thumbnails: Improved APNG detection for thumbnail generation
- Tokenization: Token estimation now uses byte length of a string
- Gallery: Added a shortcut to extensions ("Magic Wand") menu.
- a11y: Added support for system-wide reduced motion and high contrast preferences
- Reasoning: Added a new "Think XML" template without newlines in tags
- Server: Added isomorphic-git as an alternative backend for installing extensions
- Server: Added optional gzip compression for large request payloads
- Server: Added config for disallowing full data backups
- Server: Clear error messaging for port conflicts during startup
- Server: Auto-create data root in standalone mode if not exists
- Build: Improved webpack cache management to prevent stale caches
Macros
- New macro engine enabled by default for new installs
- Added:
{{charFirstMessage}}/{{greeting}}(with optional indexing for alternate greetings){{maxContextTokens}},{{maxResponseTokens}}{{allChatRange}}: provides the range of the entire chat for commands that accept ranges
- Fixed
{{summary}}prioritization of UI memory content
STscript
New Commands
- Character management:
/char-create,/char-update,/char-get,/char-delete,/char-duplicate,/tag-import - Generation control:
/swipe,/regenerate - Reasoning block control:
/reasoning-collapse,/reasoning-expand,/reasoning-toggle - Array utilities:
/array-wrap,/array-unwrap - Loader:
/loader-wrap,/loader-show,/loader-hide,/loader-stop
Enhancements
- Extended
/input,/popup,/buttonswith placeholders, tooltips, icons /imaginenow supportsgalleryargument/bgcolrebuilt for full theme palette generation using color theory
Removals
- Removed deprecated commands:
/lockand/bind(aliased to/persona-lock) - Deprecated
formatparameter for message-sending commands (usereturn)
Extensions
For developers
- Added extension lifecycle hooks via manifest declarations
- Exported
SlashCommandEnumValueandgenerateRawData - Extensions can now process streamed text during tool-call chains
- Popup System:
- Added textarea support
- Added placeholders, tooltips, icons
- Action Loader:
- New reusable loading overlay system
- Supports stacking, toasts, and STscript integration
- Added events:
PERSONA_CHANGEDTTS_JOB_STARTEDTTS_AUDIO_READYTTS_JOB_COMPLETE
Vector Storage
- Ollama: Migrated to
/api/embedwith batch support - SiliconFlow: Added as embeddings provider
- Google: Added gemini-embedding-2-preview
Image Generation
- Overriden image dimensions via the slash command are preserved for swiped images
- Manual prompt refinement: edit or discard saved negative prompt and resolution
Bug Fixes
- Fixed sanitization of chat import names
- Fixed secrets being present in data backups when not intended
- Fixed user data paths validation in route handlers
- Fixed Claude multipart text block parsing (non-streaming)
- Fixed tokenizer requests requiring model name (llama.cpp)
- Fixed embedding requests for custom API paths (llama.cpp, vLLM)
- Fixed stop string message clean-up for Chat Completion
- Fixed OpenRouter cache-at-depth logic to exclude system messages
- Fixed OpenRouter not respecting
enableThoughtSignaturessetting - Fixed OpenRouter payload field name (
input_video→video_url) - Fixed KoboldCpp DRY sequence stringification
- Fixed KoboldCpp and Tabby tokenization counting a BOS token
- Fixed Escape key closing non-dismissable popups (double tap Escape to override)
- Fixed tag visibility persistence in Tag Management
- Fixed priority of closing side panels with Escape keys
- Fixed APNG image format detection in thumbnail handling
- Fixed duplicate image-generation toasts
- Fixed tool-call error handling and reporting
- Fixed typo in KoboldAI instruct template (
OUPUT→OUTPUT)
Community updates
- Chore: Add code formatting conventions as eslint rules by @Cohee1207 in #5158
- Chore: enable brace-style eslint check by @Cohee1207 in #5159
- Fix/Remove non-samplers from sampler selection by @leandrojofre in #5163
- Fix NanoGPT prompt caching for anthropic/claude-* model IDs by @SenatusSPQR1 in #5164
- Overhaul /bgcol: Oklch color space, full palette generation, save as new theme by @Copilot in #5162
- Fix koboldcpp dry sequence stringification by @malformed-c in #5156
- feat: add prefers-reduced-motion and prefers-contrast media queries by @xavisavvy in #5122
- Add Textarea Support to Popup Custom Inputs by @Wolfsblvt in #5167
- Guard
isValidImageUrlagainst nullishavatar_urlinputs by @Copilot in #5172 - Guard
isValidImageUrlagainst nullishavatar_urlinputs (#5172) by @Cohee1207 in #5173 - i18n: Comprehensive update and refactor for Simplified Chinese (zh-CN) by @YKDZ in #5171
- Add gallery argument to /imagine command by @bal-spec in #5168
- Add {{maxContext}}, {{maxResponse}} macros and {{maxPromptTokens}} alias by @Copilot in #5176
- Add gemini-3.1-pro-preview to Google AI Studio and Vertex model lists… by @SpicyMarinara in #5188
- Add gemini-3.1-pro-preview to caption settings by @SpicyMarinara in #5190
- Update zh-CN translations by @Tosd0 in #5180
- Fix APNG thumbnail handling and improve animated format detection by @Vibecoder9000 in #5113
- Feat/Export SlashCommandEnumValue for extensions through getContext by @leandrojofre in #5191
- build llama.cpp embedding url path by appending instead of overwriting by @wbruna in #5178
- OpenRouter interleaved reasoning forwarding for tool-call continuations by @ZhenyaPav in #5160
- Preserve image dimensions when generating swipe images by @Copilot in #5205
- feat(openrouter): disable reasoning if Request model reasoning is off and effort is minimum by @Brioch in #5079
- Open router provider filter by @GERNOMA in #5208
- Add PERSONA_CHANGED event by @Copilot in #5218
- Background Folders by @Vibecoder9000 in #5187
- Fix tag hide/show toggle in Tag Management not persisting by @Copilot in #5226
- Use Ollama /api/embed endpoint for vector embeddings by @shifusen329 in #5221
- Remove open bogus folder tags that don't have any entities assigned by @Cohee1207 in #5225
- Add unit tests for pure functions in src/util.js by @tgies in #5231
- fix: skip system messages in OpenRouter cacheAtDepth calculation by @tgies in #5230
- Support for isomorphic-git as an alternative git backend, part 1 by @Sanitised in #5229
- Claude: map Reasoning Effort to adaptive thinking config by @Cohee1207 in #5219
- Implement {{charFirstMessage}} / {{greeting}} macro with alternate greeting indexing and substitution by @Copilot in #5220
- Enable experimental macro engine by default (for new users) by @Wolfsblvt in #5234
- Add gpt-5.3-chat-latest model support by @SpicyMarinara in #5241
- Remove xAI web search capability by @equal-l2 in #5255
- Group add to background folder by @Vibecoder9000 in #5237
- add Slash commands regenerate and swipe by @hcsum in #5243
- Split generateRaw into two functions exposing generateRawData by @Kristyku in #5249
- Implement extension manifest hooks for lifecycle events by @Copilot in #5261
- feat: optionally gzip large save uploads with fallback by @Roland4396 in #5259
- Use string byte length for token guesstimation by @Cohee1207 in #5267
- fix: prioritize UI memory content for {{summary}} macro resolution by @Pavdig in #5268
- feat: Add gemini-embedding-2-preview in google vectorization models by @MakksSh in #5274
- feat: add array-wrap and array-unwrap commands by @Cohee1207 in #5277
- fix: update non-streaming parsing for Claude multi-parts by @Cohee1207 in #5278
- Fix typo in KoboldAI.json: "OUPUT" → "OUTPUT" by @Copilot in #5282
- Chat lorebook click unify by @Reisenbug in #5246
- Add reasoning-collapse, reasoning-expand, reasoning-toggle slash commands with range support by @Copilot in #5296
- ci: Split PR checks and PR manager workflows by @Cohee1207 in #5300
- Intellectual Webpack cache management by @Cohee1207 in #5295
- Fix AICC direct link import parsing by @GentleBurr in #5307
- Character CRUD Slash Commands by @Wolfsblvt in #5306
- fix(openrouter): respect gemini.thoughtSignatures setting for message … by @DN2331 in #5318
- Remove long deprecated STscript commands and parameters by @Wolfsblvt in #5312
- feat(tts): emit events and track messageId for third-party integrations by @xz-dev in #5309
- Add Action Loader Module with Stacking Support and STscript Commands by @Wolfsblvt in #5311
- Extend Popup System with Placeholder, Tooltip, and Icon Support (+ those in slash commands) by @Wolfsblvt in #5322
- fix: remove deprecated xAI grok-2-image-1212 model by @xz-dev in #5334
- Correct input_video to video_url in embedOpenRouterMedia() by @cloak1505 in #5331
- Fix: Prevent Escape Key from Closing Non-Dismissable Popups by @Wolfsblvt in #5327
- feat: add SiliconFlow.cn chat completion and embedding support by @xz-dev in #5316
- Filter out inapplicable stop strings for Chat Completion by @DeclineThyself in #5337
- Deprecate legacy loader and migrate all callsites to action-loader system with informative toasts by @Wolfsblvt in #5326
- Fix missing model name in tokenize requests for llama.cpp (fixes #4962) by @allo- in #5344
- Add
allowEscapeCloseoption to popup class for Escape key behavior override by @Wolfsblvt in #5340 - Handle port conflicts during server startup by @Syferus in #5349
- Renaming a lorebook now re-links itself to cards using it (with a confirmation prompt) by @egirlmilkers in #5323
- Fix vLLM vector embeddings URL construction to preserve custom API path prefixes by @Copilot in #5350
- Fix npm audit (yaml, picomatch) by @Cohee1207 in #5354
- feat: add finalizeIntermediaryMessage by @xz-dev in #5333
- feat(ui): add interface to manage message swipe history by @awaae001 in #5304
- CI: SHA hash pin dependencies in Actions by @Cohee1207 in #5358
- Update release version to 1.17.0 and fix audit in dependencies by @Cohee1207 in #5359
- fix: conditionally include secrets in user data backup by @Cohee1207 in #5360
- Add glm-5.1 to models list by @Cohee1207 in #5361
- Fix async file deletion bugs in assets endpoint by @tgies in #5363
- 1.17.0 by @Cohee1207 in #5368
New Contributors
- @SenatusSPQR1 made their first contribution in #5164
- @malformed-c made their first contribution in #5156
- @xavisavvy made their first contribution in #5122
- @YKDZ made their first contribution in #5171
- @SpicyMarinara made their first contribution in #5188
- @wbruna made their first contribution in #5178
- @GERNOMA made their first contribution in #5208
- @shifusen329 made their first contribution in #5221
- @tgies made their first contribution in #5231
- @Sanitised made their first contribution in #5229
- @hcsum made their first contribution in #5243
- @Roland4396 made their first contribution in #5259
- @MakksSh made their first contribution in #5274
- @Reisenbug made their first contribution in #5246
- @GentleBurr made their first contribution in #5307
- @DN2331 made their first contribution in #5318
- @xz-dev made their first contribution in #5309
- @allo- made their first contribution in #5344
- @Syferus made their first contribution in #5349
- @egirlmilkers made their first contribution in #5323
Full Changelog: 1.16.0...1.17.0
SillyTavern 1.16.0
Note: The first-time startup on low-end devices may take longer due to the image metadata caching process.
Backends
- NanoGPT: Enabled tool calling and reasoning effort support.
- OpenAI (and compatible): Added audio inlining support.
- Added Adaptive-P sampler settings for supported Text Completion backends.
- Gemini: Thought signatures can be disabled with a config.yaml setting.
- Pollinations: Updated to a new API; now requires an API key to use.
- Moonshot: Mapped thinking type to "Request reasoning" setting in the UI.
- Synchronized model lists for Claude and Z.AI.
Features
- Improved naming pattern of branched chat files.
- Enhanced world duplication to use the current world name as a base.
- Improved performance of message rendering in large chats.
- Improved performance of chat file management dialog.
- Groups: Added tag filters to group members list.
- Background images can now save additional metadata like aspect ratio, dominant color, etc.
- Welcome Screen: Added the ability to pin recent chats to the top of the list.
- Docker: Improved build process with support for non-root container users.
- Server: Added CORS module configuration options to config.yaml.
Macros
Note: New features require "Experimental Macro Engine" to be enabled in user settings.
- Added autocomplete support for macros in most text inputs (hint: press Ctrl+Space to trigger autocomplete).
- Added a hint to enable the experimental macro engine if attempting to use new features with the legacy engine.
- Added scoped macros syntax.
- Added conditional
ifmacro and preserve whitespace (#) flag. - Added variable shorthands, comparison and assignment operators.
- Added
{{hasExtension}}to check for active extensions.
STscript
- Added
/reroll-pickcommand to reroll{{pick}}macros in the current chat. - Added
/beepcommand to play a message notification sound.
Extensions
- Added the ability to quickly toggle all third-party extensions on or off in the Extensions Manager.
- Image Generation:
- Added image generation indicator toast and improved abort handling.
- Added stable-diffusion.cpp backend support.
- Added video generation for Z.AI backend.
- Added reduced image prompt processing toggle.
- Added the ability to rename styles and ComfyUI workflows.
- Vector Storage:
- Added slash commands for interacting with vector storage settings.
- Added NanoGPT as an embeddings provider option.
- TTS:
- Added regex processing to remove unwanted parts from the input text.
- Added Volcengine and GPT-SoVITS-adapter providers.
- Image Captioning: Added a model name input for Custom (OpenAI-compatible) backend.
Bug Fixes
- Fixed path traversal vulnerability in several server endpoints.
- Fixed server CORS forwarding being available without authentication when CORS proxy is enabled.
- Fixed asset downloading feature to require a host whitelist match to prevent SSRF vulnerabilities.
- Fixed basic authentication password containing a colon character not working correctly.
- Fixed experimental macro engine being case-sensitive when checking for macro names.
- Fixed compatibility of the experimental macro engine with the STscript parser.
- Fixed tool calling sending user input while processing the tool response.
- Fixed logit bias calculation not using the "Best match" tokenizer.
- Fixed app attribution for OpenRouter image generation requests.
- Fixed itemized prompts not being updated when a message is deleted or moved.
- Fixed error message when the application tab is unloaded in Firefox.
- Fixed Google Translate bypassing the request proxy settings.
- Fixed swipe synchronization overwriting unresolved macros in greetings.
Community updates
- Macros 2.0 (v0.4) - Add scoped macros (last arg can be scoped),
{{if}}macro and macro flags (baseline implementation) by @Wolfsblvt in #4913 - Zai moonshot reverse proxy by @subzero5544 in #4923
- add new Tts adapter provider by @guoql666 in #4915
- updated claude prompt caching url by @underscorex86 in #4931
- Macros 2.0 (v0.5) - Add variable shorthand macros and variable support to
{{if}}macro by @Wolfsblvt in #4933 - Added regex filter option to TTS extension by @ZhenyaPav in #4924
- Macros 2.0 [Fix] - Make macro name matching case-insensitive throughout the macro system by @Wolfsblvt in #4942
- Macros 2.0 (v0.5.1) - Delayed Macro Argument Resolution for
{{if}}Macro by @Wolfsblvt in #4934 - Add Adaptive-P settings by @Cohee1207 in #4945
- Wrap reloadCurrentChat into SimpleMutex by @Cohee1207 in #4944
- Macros 2.0 [➕Macros] - Add
{{hasExtension}}macro, and refactor extension lookup logic by @Wolfsblvt in #4948 - Macros 2.0 [Fix] - Macro display override uses alias name if relevant + preserve whitespaces for autocomplete closing tag by @Wolfsblvt in #4953
- Macros 2.0 (v0.5.2) - Onboarding for new Macro Engine by @Wolfsblvt in #4955
- Caption: Add custom model input field by @Cohee1207 in #4956
- Macros 2.0 (v0.6.6) - STscript compatibility by @Cohee1207 in #4957
- Audio inlining for OpenAI and Custom-compatible by @Cohee1207 in #4964
- Removed 87 redundant
chat[chat.length - 1]lookups. by @DeclineThyself in #4963 - Fix user handle naming logic by @allen9441 in #4969
- "gradually replacing property access with a dot operator" by @DeclineThyself in #4965
- Optimize getGroupPastChats by @Cohee1207 in #4976
- Replace $.ajax with fetch by @Cohee1207 in #4978
- Macros 2.0 [Fix] - Fix macro evaluation to allow nested scoped macros in arguments by @Wolfsblvt in #4977
- Macros 2.0 [Fix] - Fix
{{pick}}macro seeding inside delayed-resolution macros like{{if}}by @Wolfsblvt in #4986 - Fix: init macros before extensions by @Cohee1207 in #4988
- Enhance world duplication to use current world name as base by @Wolfsblvt in #4990
- Fix: don't call append media twice on swipe by @Cohee1207 in #4991
- Improve performance of printMessages by @Cohee1207 in #4979
- Feature: Enhanced Branch and Checkpoint Naming by @Wolfsblvt in #4993
refactor/perf-printMessages#2: RemovedgetMessageFromTemplateby @DeclineThyself in #4983- Show a page reload prompt on EME toggle by @Cohee1207 in #4994
refactor/perf-printMessages#3: ExtractedupdateMessageItemizedPromptButtonandgetMessageHTMLfromaddOneMessageto improve readability. by @DeclineThyself in #4984- Adjust itemized prompts on message move/delete by @Cohee1207 in #5000
- Macros 2.0 (v0.7.0) -Variable Shorthand: New Operators & Lazy Evaluation by @Wolfsblvt in #4997
refactor/perf-printMessages#4: RenamednewMessagetomessageElementandnewMessageIdtomessageId. by @DeclineThyself in #4985- Add to list of Showdown block tags by @Cohee1207 in #4998
- Update Dockerfile by @Cohee1207 in #4954
- Suppress error messages when Firefox unloads the tab by @Cohee1207 in #5013
- feat(sd): Add Z.AI GLM-Image model support by @mschienbein in #5012
- Improved
printMessagesperformance on large chats by reducing DOM updates. by @DeclineThyself in #4947 - Adaptive-P for llama.cpp llama-server by @Beinsezii in #4959
- Add taxon filter controls to Group Chat member list by @paradox460 in #5006
- Adaptive P Hotfix by @Beinsezii in #5022
- Adding Slash Commands for Vector Storage Extension by @adventchilde in #5008
- Gemini: Add config.yaml setting for thought signatures by @Cohee1207 in #5025
- Docker: Build Optimization and Enhanced Non-Root/Volumeless Support by @Pavdig in #5024
- feat(sd): Add generation status indicator and improve abort handling by @mschienbein in #5015
- Recent Chats: Add pin functionality by @Cohee1207 in #5030
- feat(docker): add robust healthcheck script by @Cohee1207 in #5028
- Fixes #4950:try to modify config.yaml at start instead of modify it by @san-tian in #5043
- adaptive_P for tabby by @Ph0rk0z in #5044
- Add
APP_INITevent before hideLoader in initialization sequence, beforeAPP_READYfires by @Wolfsblvt in #5051 - Add
/reroll-pickcommand to reset{{pick}}macro by @Wolfsblvt in #5049 - Added
'dot-notation': ['error']to.eslint.cjsby @DeclineThyself in #5042 - Clarified contribution guidelines for large PRs. by @DeclineThyself in #5032
- Allow reasoning edit to substitute macros on saving the reasoning by @Wolfsblvt in #5052
- Macros 2.0 (v0.7.3) - Variable Shorthand: Comparison Operators & Autocomplete Improvements by @Wolfsblvt in #5050
- Macros 2.0 - [Chore] Allow registration of aliases for existing macros by @Wolfsblvt in #5053
- Macros 2.0 (v0.7.1) - Macro Autocomplete everywhere by @Wolfsblvt in #5019
- Update Pollinations API by @Cohee1207 in #5060
- Expose character update APIs for extensions by @rdeforest in #5062
- Volcengine tts by @Crush0 in #5003
- Stable diffusion.cpp server support by @Jay4242 in #5074
- fix: welcome depth by @StageDog in #5077
- feat(openrouter): add model quantizations setting by @Brioch in #5080
- Refactor /search to use per-line async parsing by @Cohee1207 in #5085
- /image-metadata by @Vibecoder9000 in #4788
- Fix/Do not spam saveSettingsDebounced in AccountStorage by @leandrojofre in #5090
- Macros 2.0 - Improve Autocomplete edge cases on completing macros by @Wolfsblvt in #5093
- (chore) World Info slash commands do some console context warn loggings by @Wolfsblvt in #5096
- Add Minimal Prompt Processing option by @KrsityKu in #5095
- add option for claude-opus-4-6 by @LumiWasTaken in #5103
- Feat/Allow to bulk toggle all third-party extensions from Manage Extensions by @leandrojofre in #5094
- Backgrounds metadata population and frontend colors by @Vibecoder9000 in #5092
- Add
clearDataoption toclearChatfunction by @DeclineThyself in #5091 - Sync OpenRouter providers list by @cloak1505 in #5110
- Background sort feature by @Vibecoder9000 in #5107
- Macros 2.0 - Optional scoped content + improved closing-tag autocomplete by @Wolfsblvt in #5117
- fix: sync swipes only when chat is not pristine to ensure macro resolution by @Cohee1207 in #5106
- fix: correct typo 'seperated' to 'separated' by @thecaptain789 in #5121
- Add rename buttons for ComfyUI workflows and style presets by @Copilot in #5124
- feat(server): make CORS middleware configurable by @awaae001 in #5123
- Preserve user input on tool call recursion by @Cohee1207 in #5134
- Set HTML lang attribute from app locale to enable CSS hyphens: auto by @Copilot in #5136
- Add "✨ Vibe Coded" label to PRs created by @Copilot by @Copilot in #5137
- Add GLM-5 to Z.AI model list by @Copilot in #5138
- Macros 2.0 -
list-supported Macros Autocomplete Improvements by @Wolfsblvt in #5135 - Update zh-CN translations by @Tosd0 in #5145
- Add NanoGPT embeddings support for Vector Storage by @Copilot in #5150
- Fix: HTTP Basic Auth fails when password contains colons by @Hime-Hina in #5153
- Staging by @Cohee1207 in #5154
New Contributors
- @guoql666 made their first contribution in #4915
- @ZhenyaPav made their first contribution in #4924
- @allen9441 made their first contribution in #4969
- @mschienbein made their first contribution in #5012
- @paradox460 made their first contribution in #5006
- @adventchilde made their first contribution in #5008
- @san-tian made their first contribution in #5043
- @rdeforest made their first contribution in #5062
- @Crush0 made their first contribution in #5003
- @Jay4242 made their first contribution in #5074
- @Brioch made their first contribution in #5080
- @KrsityKu made their first contribution in #5095
- @thecaptain789 made their first contribution in #5121
- @Copilot made their first contribution in #5124
- @Hime-Hina made their first contribution in #5153
Full Changelog: 1.15.0...1.16.0
SillyTavern 1.15.0
Highlights
Introducing the first preview of Macros 2.0, a comprehensive overhaul of the macro system that enables nesting, stable evaluation order, and more. You are encouraged to try it out by enabling "Experimental Macro Engine" in User Settings -> Chat/Message Handling. Legacy macro substitution will not receive further updates and will eventually be removed.
Breaking Changes
{{pick}}macros are not compatible between the legacy and new macro engines. Switching between them will change the existing pick macro results.- Due to the change of group chat metadata files handling, existing group chat files will be migrated automatically. Upgraded group chats will not be compatible with previous versions.
Backends
- Chutes: Added as a Chat Completion source.
- NanoGPT: Exposed additional samplers to UI.
- llama.cpp: Supports model selection and multi-swipe generation.
- Synchronized model lists for OpenAI, Google, Claude, Z.AI.
- Electron Hub: Supports caching for Claude models.
- OpenRouter: Supports system prompt caching for Gemini and Claude models.
- Gemini: Supports thought signatures for applicable models.
- Ollama: Supports extracting reasoning content from replies.
Improvements
- Experimental Macro Engine: Supports nested macros, stable evaluation order, and improved autocomplete.
- Unified group chat metadata format with regular chats.
- Added backups browser in "Manage chat files" dialog.
- Prompt Manager: Main prompt can be set at an absolute position.
- Collapsed three media inlining toggles into one setting.
- Added verbosity control for supported Chat Completion sources.
- Added image resolution and aspect ratio settings for Gemini sources.
- Improved CharX assets extraction logic on character import.
- Backgrounds: Added UI tabs and ability to upload chat backgrounds.
- Reasoning blocks can be excluded from smooth streaming with a toggle.
- start.sh script for Linux/MacOS no longer uses nvm to manage Node.js version.
STscript
- Added
/message-roleand/message-namecommands. /api-urlcommand supports VertexAI for setting the region.
Extensions
- Speech Recognition: Added Chutes, MistralAI, Z.AI, ElevenLabs, Groq as STT sources.
- Image Generation: Added Chutes, Z.AI, OpenRouter, RunPod Comfy as inference sources.
- TTS: Unified API key handling for ElevenLabs with other sources.
- Image Captioning: Supports Z.AI (common and coding) for captioning video files.
- Web Search: Supports Z.AI as a search source.
- Gallery: Now supports video uploads and playback.
Bug Fixes
- Fixed resetting the context size when switching between Chat Completion sources.
- Fixed arrow keys triggering swipes when focused into video elements.
- Fixed server crash in Chat Completion generation when invalid endpoint URL passed.
- Fixed pending file attachments not being preserved when using "Attach a File" button.
- Fixed tool calling not working with deepseek-reasoner model.
- Fixed image generation not using character prefixes for 'brush' message action.
Community Updates
- Gallery: Add video uploads to gallery by @Cohee1207 in #4796
- Add new WORLDINFO_SCAN_DONE event with mutable state for extensions by @Wolfsblvt in #4797
- Remove nvm install from start.sh by @Cohee1207 in #4804
- TC: Add a toggle for empty json schemas by @Cohee1207 in #4807
- /api-url: Add VertexAI region management by @Cohee1207 in #4808
- Backgrounds: Restore drawer title header by @Cohee1207 in #4809
- Add credit to bryc for writing getStringHash. by @DeclineThyself in #4811
- Refactor loadOpenAISettings by @Cohee1207 in #4815
- Fix resetting the context size when switching between CC sources by @Cohee1207 in #4816
- Update group chat metadata format by @Cohee1207 in #4805
- Trigger CHARACTER_RENAMED_IN_PAST_CHAT for group chats by @leandrojofre in #4818
- Unify chat timestamps format by @Cohee1207 in #4806
- Backport
feat/chat-treeand fix #4709 by @DeclineThyself in #4712 - Fix - Check for samplers in the connections pannel by @leandrojofre in #4822
- Prompt Manager: Make main/PHI/aux injectable by @Cohee1207 in #4829
- Chat Completion: Reduce number of toggles in AI Response Configuration by @Cohee1207 in #4821
- Empty
swipesare not handled byensureSwipes. by @DeclineThyself in #4828 - Return character prefixes to brush image generation by @drake1138 in #4833
- Backfill missing swipe_info by @Cohee1207 in #4831
- Add verbosity control by @Cohee1207 in #4837
- Vertexaisearch by @mightytribble in #4834
- Enhanced CharX Import with Asset Extraction by @axAilotl in #4825
- Gemini: Add image request settings by @Cohee1207 in #4838
- Backported
/testsfrommacros-2.0. by @DeclineThyself in #4842 - Backgrounds menu tabs by @Cohee1207 in #4845
- Sync OpenRouter providers list by @cloak1505 in #4846
- Convert OAI tool_choice to Gemini functionCallingConfig for Gemini requests by @mightytribble in #4840
- Added MockServer class for tests by @DeclineThyself in #4843
- Chutes integration by @cxmplex in #4844
- Add optional Setter to changeMainApi by @SammCheese in #4853
- Feat: Add toggle to exclude think/reason blocks from smooth streaming by @Dakraid in #4849
- Facillitate extension use of ConnectionManagerRequestService by @qvink in #4841
- custom-request: Pass api-url for Z.AI and Vertex and fix if omitted by @Cohee1207 in #4859
- Fix Mistral's Max Temperature by @kashmirmydon in #4856
- Fix path to sprites construction by @Cohee1207 in #4860
- Implement chat backups browse menu by @Cohee1207 in #4862
- Regex cache by @Cohee1207 in #4858
- Fixed
isModifiedKeyboardEventorder of operations. by @DeclineThyself in #4866 - "N" support for llama.cpp by @Beinsezii in #4869
- Backported refactor of the chats endpoint from
feat/chat-tree.by @DeclineThyself in #4870 - Add docker data directories to .dockerignore by @equal-l2 in #4873
- Add an explicit
cache_controlto the first system message for OpenRouter Claude by @chungchandev in #4872 - Fix: preserve attached files during file input changes by @Cohee1207 in #4877
- Replace Google Translate library by @Cohee1207 in #4884
- Refactor CC API async route handlers by @Cohee1207 in #4885
- Chevrons can overlap with other elements. by @DeclineThyself in #4878
- Correct error message in evalBoolean by @Kexus in #4889
- Macros 2.0 (v0.3) - Replacing the existing Macro System with a new Macro Engine by @Wolfsblvt in #4820
- Z.AI: Add image generation and web search by @Cohee1207 in #4895
- Gemini: add media resolution select by @Cohee1207 in #4775
- Implement Gemini thought signatures by @mightytribble in #4886
- Migrate substituteParams calls to new engine by @Cohee1207 in #4901
- Refactor ElevenLabs TTS API key handling by @Cohee1207 in #4906
- Add caching system prompt feature for OpenRouter Gemini by @chungchandev in #4903
- Add model selection support for llama.cpp router mode by @my-alt-acct in #4910
- Fix context size limitation for llama.cpp router mode by @my-alt-acct in #4914
- Updated Avatar Style and
Hide Chat Avatarstitle text. by @DeclineThyself in #4908 - Comfyui serverless runpod image generation by @9nbf7c4q6b-lgtm in #4891
- chchar slash cmd by @AphidGit in #4916
- filter out models that don't have a valid id for chutes by @cxmplex in #4920
- Fix OpenRouter compatibility for OpenAI models by @equal-l2 in #4917
- Allow editing of global Worldinfo settings by @SammCheese in #4921
- [Electron Hub] Prompt Caching Support for Claude models by @snowby666 in #4918
- Staging by @Cohee1207 in #4925
New Contributors
- @drake1138 made their first contribution in #4833
- @mightytribble made their first contribution in #4834
- @axAilotl made their first contribution in #4825
- @cxmplex made their first contribution in #4844
- @SammCheese made their first contribution in #4853
- @kashmirmydon made their first contribution in #4856
- @chungchandev made their first contribution in #4872
- @Kexus made their first contribution in #4889
- @my-alt-acct made their first contribution in #4910
- @9nbf7c4q6b-lgtm made their first contribution in #4891
- @AphidGit made their first contribution in #4916
Full Changelog: 1.14.0...1.15.0
@ Every self-hosted release, weekly
One Monday email with the week's stable releases across the directory.
FAQ
Can I run SillyTavern on Android?
Yes, SillyTavern can run natively on Android using Termux from F-Droid or GitHub releases.
Do I need a powerful PC to run SillyTavern?
The frontend runs on any device with Node.js 20+, but local LLM inference requires a GPU with at least 6GB VRAM.
What is SillyTavern?
SillyTavern is a user interface for interacting with AI language models, allowing you to create characters, chat with them, and manage conversation history.
More in AI Chat
SiloSilo is a self-hosted, open-source pure front-end app for multi-model AI chat and text-to-image generation with PWA and browser extension.
SmarterRouterSelf-hosted AI gateway that automatically routes queries to the best local or cloud LLM based on prompt analysis, with VRAM management and OpenAI-compatible API.
sqlchatOpen-source, self-hosted chat-based SQL client that lets developers query and edit MySQL, PostgreSQL and MSSQL databases in plain language using OpenAI or Ollama.
Telegram LLM BotSelf-hosted Telegram bot running local LLMs via Ollama, transcribing voice with faster-whisper, tool calling, web search, SQLite history, optional MinIO voice archive.
Text Generation WebUIAn open-source, self-hosted LLM chat app you run on your own machine: vision input, tool-calling, LoRA fine-tuning, and an OpenAI/Anthropic-compatible API.
ThesysSelf-hosted open-source research workspace for searching papers, uploading PDFs, on-PDF highlights with citations, reader mode Q&A, paper comparison, and citation export.
ThoughtDAGThoughtDAG is an open-source, self-hostable AI chat canvas: every exchange is a graph node, and its wires decide the context your model receives next.
Tidal EchoSelf-hosted, open-source private chat connecting a mobile PWA to your local AI via your own relay server — no accounts, no third-party hosting.
tiledeskSelf-hosted open-source live chat with AI chatbots, multichannel support, visual design studio, RAG knowledge base, and human-in-the-loop escalation.
VaneSelf-hosted, open-source AI answering engine: searches via bundled SearxNG and answers with cited sources, using local Ollama models or OpenAI, Claude and Gemini.
voice-chat-aiSelf-hosted open-source voice chat app for talking to AI characters in a web UI or terminal, mixing Ollama and cloud LLMs with Spark-TTS, Kokoro or ElevenLabs voices.
VoicEraVoicEra is an open-source, self-hosted platform for real-time phone agents combining STT, LLMs, TTS, and telephony in a Docker stack with 26+ AI providers.
0 comments