Venice.ai Change Log - February 3rd, 2025

App UI

• Inference - Update error handling on document upload to gracefully handle display of invalid PDF errors.

• Characters - Update the share URL within the character settings screen to use the character's public slug vs. UUID.

• With Enter Submits Chat disabled, permit sending the chat with control-enter. Solves request from user in Featurebase

• Remove Retiring Soon tag from Dolphin. Our intent was to retire this model and replace it with an upcoming Dolphin release but until we have a final ETA from Dolphin, the model will remain.

• Add a spinner to the Thinking... block in Dolphin to make the UI more clear that the LLM is generating content behind the scenes.

• Fixed a bug that made the Copy option on code blocks not possible to click until the entire message completed rendering.

Token Dashboard

• Add a key to the Network Utilization Graph

• Fix rendering of VCU cards on mobile screens

• Force wallets to connect to the Base network when executing transactions on-chain.

• Add "Claim and Restake" button to facilitate claiming and immediately restaking rewards in a single transaction.

• Create Dune Analytics dashboard showing network utilization over time as recorded on-chain.

API

• Fixed issue where the use of max_completion_tokens in combination with the llama-3.1-405b model would result in a 500 response.

• Support light and dark mode, toggle-able in the top right corner.