Changelog

Text models3

• Llama 3B

  Added support for web search to Llama 3B.

  All users

• Mistral 24B

  Mistral 24B is now the default Venice model for all users, bringing multi-modal capabilities to everyone.

  All users

• Llama 70B

  Llama 70B has been updated to a Pro model.

  Pro users

Image models4

• Upscale & Enhance

  A new upscaling architecture with proprietary enhancement that creatively in-fills new pixels at higher resolution. Offers two modes: Upscale for accurate enlargements and Enhance for adding new details with adjustable creativity.

  Pro users

• Venice SD35

  Venice SD35 is the new default image model, a custom-configured Stable Diffusion 3.5 engine powered by a Comfy UI workflow backend. It improves natural language understanding and delivers superior results through an optimized custom pipeline.

  All users

• Image Inference Infrastructure

  Launched a preview of the new image inference infrastructure to beta testers.

  Beta

• Default Image Model

  Changed default image model to venice-sd35 to match app default.

  All users

Chat11

• Bulk Chat Deletion

  Added support for bulk chat deletion. Select the three dots to the right of the chats header to find the option.

• Character Ratings

  Added character ratings to public Venice characters, allowing users to rate and help curate the best characters on the platform.

• Conversation Forking

  Added Conversation Forking to allow splitting conversations in new directions.

• Delete History Progress Bar

  Added a progress bar to the UI when deleting chat history, providing clearer feedback for users with large histories.

• Edit Conversation History

  Pro members can now edit their conversation history, a highly requested feature with 371 votes on Featurebase.

  Pro users

• Enter to Save Conversation Name

  Support pressing Enter to save the name when re-naming a conversation.

• Show 5 more

Image, video and audio2

• Favorite None Image Style

  Added support to allow the “None” image style to be added to favorites.

• Searchable Image Styles

  Image styles within the image settings are now searchable.

Library and media2

• Character Search Delay

  Added additional delay to the character search to avoid re-rendering while search results are in flight.

• Recent Activity Sorting

  Folders and characters in the left navigation bar are now sorted by recent activity.

Account and settings1

• Authentication Token Reload

  Updated authentication logic to programmatically reload authentication tokens if they time out due to the window being backgrounded.

Wallet and payments1

• Token Dashboard Precision

  Update token dashboard to show 2 digits of precision for staked VVV.

General7

• External Link Warning Setting

  Added a setting in the App settings panel to disable the warning on external links.

• New Backend Service

  Launched a new backend service to support the growth of Venice. Migrated authentication, image generation and upscale to that service for better performance and scalability.

• Upscale Enhance Mode

  Venice's new upscale enhance mode is now available through the existing upscale API.

• Character Moderation System

  Revised the Character moderation system to reduce time to approvals.

• Code Block Styling

  Updated code block and inline code styling for improved readability.

• Native Confirmation Windows

  Swapped out browser confirmation windows with native confirmation windows to allow them to be displayed on mobile crypto wallet browsers like Coinbase Wallet.

• Show 1 more

Mobile app2

• Character Page Navigation

  Updated navigation on character pages making them easier to navigate on mobile.

• Mathjax Horizontal Scroll

  Added support for horizontal scroll in Mathjax containers on mobile.

• response_format for Qwen VL

  Added response_format support to Qwen VL in the API.

• JSON Object Response Format

  Added support for json_object type under response_format for LLM API requests.

• nextEpochBegins in Rate Limits

  Added the nextEpochBegins key to the api_keys/rate_limits endpoint. Documentation has been updated.

• OpenAI-Compatible Image Endpoint

  Added an OpenAI compatible image generation endpoint. This endpoint supports fewer configuration parameters than our full image endpoint but works with OpenAI compatible image generation libraries without modification.

• Improved LLM Error Messages

  Return more helpful errors when LLM API requests fail to execute on the runners. This should make debugging complex json_schemas easier to understand.

• New Backend Migration

  All chat inference in the Venice app has been migrated to Venice's new backend infrastructure, improving performance and maintainability.

• Markdown Rendering Fix

  Updated Markdown rendering to correctly handle messages containing $ signs, fixing a bug where responses were incorrectly interpreted as LaTeX.

  Chat

• Text Settings Not Saving

  Fixed a bug with text settings not properly saving.

  Chat

• Image Variant Error Handling

  Update image variant generation to more gracefully handle error cases.

  Studio

• Wallet Connect Network Notice

  Updated Wallet Connect packages to fix a bug that was causing a notice to pop up prompting to change wallet networks.

  Wallet

• Document Whitespace Stripping

  Fixed a bug that was causing whitespace to be stripped from uploaded documents prior to LLM processing.

• Left Nav Scroll Issue

  Fixed a bug that was causing scroll issues in the left nav bar.

• Show 3 more