Changelog
API Docs
Help Center

March 30th, 2026

Venice.ai Change Log - March 7, 2026 - March 26, 2026

Venice Referral Program

Share your referral link and earn $10 in API credits when someone you refer upgrades to Pro. Referred users also receive $10 in credits. Track referrals from Profile settings or the new referral drawer. The program is open to all users, including anonymous visitors.

Privacy Modes: TEE & E2EE on Venice

Every model on Venice now displays a privacy mode, giving users full visibility into how their data is handled. Models running inside Trusted Execution Environments (TEE) ensure isolated processing, and end-to-end encrypted (E2EE) models go further — not even Venice can see your prompts or responses. Click any privacy badge to see exactly what each mode means.

Read the full announcement →

Explore Privacy Modes →

All models below support both TEE and E2EE. Toggle between modes in model settings.

  • Venice Uncensored 1.1

  • GLM 5

  • Qwen3.5 122B A10B

  • Gemma 3 27B

  • GLM 4.7

  • GLM 4.7 Flash

  • GPT OSS 20B

  • GPT OSS 120B

  • Qwen 2.5 7B

  • Qwen3 30B A3B

  • Qwen3 VL 30B A3B

Operated by NEAR AI Cloud and Phala Network.

Import Memory

Import conversation memories from ChatGPT, Claude, and other AI providers into Venice. Rolled out to all users.

https://x.com/AskVenice/status/2033989325830955458

Kling Reference-to-Video

Upload a reference image and generate video that maintains the same character across scenes. Supports character consistency and lipsync in AI-generated videos.

https://www.youtube.com/watch?v=vjA9pECWfTY

New Models

Models added to Venice during this period.

Text Models

  • MiniMax M2.7 — Text model from MiniMax with multi-turn conversation support.

  • Aion 2.0 — Text model from AionLabs. Pro only.

  • Grok 4.20 Beta — xAI's latest Grok text model with function calling and built-in X Search.

  • Grok 4.20 Multi-Agent Beta — Multi-agent variant of Grok 4.20 with X Search support.

  • Mistral Small 3.2 24B Instruct — Mistral AI's 24B-parameter instruction-tuned text model. API only.

  • Qwen 3 Next 80B — 80B parameter dense model from Alibaba Cloud. API only.

  • Qwen 3 235B A22B Instruct — 235B parameter mixture-of-experts model from Alibaba Cloud. API only.

  • Qwen 3 235B A22B Thinking — 235B parameter MoE model with extended thinking mode. API only.

Video Models

  • Kling O3 Pro Reference-to-Video — Character-consistent video generation from a reference image. Pro tier.

  • Kling O3 Standard Reference-to-Video — Character-consistent video generation. Standard tier.

  • Seedance 1.5 Pro — Video generation model from ByteDance. Text-to-video and image-to-video. Uncensored.

Audio Models

  • ElevenLabs Multilingual v2 — Multilingual text-to-speech from ElevenLabs.

  • ElevenLabs TTS v3 — Latest ElevenLabs text-to-speech model.

  • ElevenLabs Music — Music generation from ElevenLabs.

  • ElevenLabs Sound Effects — Sound effects generation from ElevenLabs.

  • Qwen 3 TTS 0.6B — Compact text-to-speech model from Alibaba Cloud.

  • Qwen 3 TTS 1.7B — Larger text-to-speech model from Alibaba Cloud.

Model Upgrades

  • DeepSeek V3.2 — Now supports function calling and structured output. Pro only.

  • Qwen 3.5 9B — Now supports vision input, accepting both text and image prompts.

App

New Features

  • Music Generation Duration — Continuous duration control for music generation using ElevenLabs models

  • Video Queue Progress Bar — Queue items with unknown progress now show an indeterminate loading state

  • Image Generation Timeout Handling — Timeout errors during image generation now display a user-facing error message instead of failing silently

  • HEIC Image Upload Support — HEIC images are now automatically converted on upload so iOS photos can be used directly

  • Drag-and-Drop for Image Edit Prompt — Images can now be dropped directly into the edit image prompt input field

  • Qwen3 TTS in Playground — Added Qwen3 text-to-speech models to the Playground model selector

  • Resolution-Based Max Duration — Video models now enforce maximum generation duration based on selected output resolution

  • Referral Drawer — Dedicated drawer component for viewing and managing referral details

  • Referrals for Anonymous Users — Anonymous users can now view and interact with referral prompts before signing up

  • Referral Pro Upgrade Date — Referral dashboard now displays the date each referred user upgraded to Pro

  • Image Info Button in Edit Tab — Info button on the selected image in the edit tab showing image metadata

  • Video Duration Ranges on Model Cards — Model cards now show supported video duration ranges

Wallet and Payments

  • Coinbase Payment Links — Added Coinbase payment links as a payment option

  • Coinbase API Migration — Migrated Coinbase payments from Commerce API to Payment Links API

  • Referral Credits in Profile — Referral credits earned now displayed in Profile settings

  • Out of Credits Flow — User flow for when credit balance reaches zero, prompting next steps

  • Bonus Credits Callout Card — New card component highlighting available bonus credits

  • Admin Referral Credits Page — New admin page to view and manage referral credits

  • Admin Referrals Endpoint — New admin API endpoint for querying user referrals and credits

  • Referral Credits in Session API — Referral credit data now included in the user session API response

  • Stripe Invoice Pagination — Added Stripe hasMore parameter to invoice list responses

  • Auto Top-Up Amount Update — "Amount to Add" field now auto-updates when Auto Top-Up is configured

  • Next Estimated Payment — Stripe users can now see their next estimated payment amount

Mobile App

  • Connection Error Message — Added a user-facing error message when the connection is dropped mid-request.

  • Document Attachment Loader — Fixed missing loading indicator when attaching documents in chat.

  • Document Pick Wallet Fix — Fixed an issue where selecting multiple files for document upload caused the wallet to log out.

  • Import Memory Prompt Localization — Added internationalization support for the import memory prompt text.

  • Android Header Fix — Fixed the right-side header element not rendering in the Android stack header.

  • Social Feed Video Playback — Fixed playback issues and improved video rendering in the social posts feed.

  • Social Feed Media Layout — Added inline video playback and a masonry grid layout for media in the Social Feed.

  • Background Removal — Added a Remove Background action for images in the mobile app.

  • Video Playback UI — Improved video player interface elements and expanded test coverage.

  • Video Loading Progress — Video loader now displays percentage complete and estimated remaining time.

  • Enhanced Prompt Text Selection — Text inside the enhanced prompt popover is now selectable.

  • Lazy-Loaded Tabs — Image and feed tabs now lazy-load their content instead of rendering on mount.

  • Filter Defaults and Reset — Adjusted default filter values and added a reset button to restore defaults.

  • Chat Memory Default — Chat memory is now enabled by default for new sessions.

  • Video Models in Top-Level Navigation — Video models now appear at the same navigation level as text and image models.

  • Android Video Tap Fix — Fixed tapping a playing video on Android not registering correctly.

  • Media Detail Pre-Loading — Images now pre-load before navigating to the media detail screen.

  • Video State on Tab Swipe — Fixed videos continuing to play when swiping between tabs.

  • Video Feed Screen — Added a dedicated video feed screen to the mobile app.

  • Image Feed Screen — Added a dedicated image feed screen to the mobile app.

  • Video Tab — Added a new Video tab to the main tab layout.

  • Prompt Screen Keyboard Dismiss — Keyboard now dismisses when leaving the prompt screen.

  • Video Feed Volume Control — Added a volume control to videos in the feed.

  • Assistant Video Error Boundaries — Wrapped assistant video components in error boundaries to prevent crashes.

  • Post Video Error Boundaries — Wrapped video components in social posts with error boundaries.

  • Prompt Screen Render Optimization — Reduced re-renders on the prompt screen to fix a Hermes runtime crash.

  • Import Memories Dark Mode — Applied dark mode styling to the Import Memories button; added translations for multiple languages.

  • Prompt Input Limits — Added character and token limits to the message input components.

  • Image Modal Orientation Fix — Fixed image modal orientation for large portrait images.

API

  • Video Transcript API — New public endpoint for video transcription at $0.02 per request.

  • Music Services v1 API — New /v1/music API endpoints for music generation.

  • Character Reviews Endpoint — New GET /v1/characters/[slug]/reviews endpoint to retrieve character reviews.

  • Characters Public API Expansion — Added previously missing fields and features to the characters public API.

  • GTM on API Settings Page — Added Google Tag Manager loader to the API settings page.

  • Character Author Field — Characters now use a dedicated author field in their data model.

  • Chat Response Time — Chat responses now display elapsed response time.

  • Anonymous API Settings Access — Unauthenticated users can now view the API settings page.

  • Characters Endpoint: author & isOwner — /characters endpoint now returns author and isOwner fields.

  • System-Only Message Requests — API now accepts requests containing only system messages.

  • Default Reasoning Effort — Default reasoning effort for inference set to medium.

Fixes and Improvements

  • Improved file-type support detection and validation in the chat document upload flow

  • Updated referral text to clarify that referred users must subscribe to Pro

  • Improved responsiveness when syncing available models with the server

  • Updated document upload panel positioning to auto-adapt to available screen space

  • Fixed a crash caused by the chat message timer exceeding maximum update depth

  • Fixed clipped content caused by overflow CSS being set to hidden

  • Fixed a crash when displaying the message header for conversations using retired chat models

  • Fixed a createTreeWalker crash on the chat page

  • Added timeout handling for Memoria long-term memory operations

  • Fixed the Learn More link on the Diem token page pointing to an incorrect destination

  • Improved how video outputs and scene images are stored and displayed within chat

  • Simplified the duration label formatting shown during video generation

  • Fixed badge rendering and layout on mobile web viewports

  • Updated memory import prompt to request markdown-formatted input

  • Improved memory import to batch embedding requests and avoid rate limits

  • Strengthened rate limiting and endpoint validation across REST API routes

  • Removed maxBatchSize field from API response payloads

  • Restored Aria voice option in text-to-speech

  • Fixed referral credits not displaying correctly in the referrals drawer

  • Fixed referral credits being incorrectly issued for referrals made after subscription

  • Fixed price helpers and removed deprecated three-day subscription type

  • Fixed Open Graph thumbnails not displaying correctly for feed posts

  • Fixed inpaint and upscaler error messages not showing properly for free-tier users

  • Fixed race condition causing image attachments to disappear from conversations

  • Fixed defects in the Video Editor

  • Improved image size error messaging

  • Added video transcription v1 endpoint with fixed pricing

  • Improved chat interface to render sub-agent steps and responses

  • Fixed an error on the chat page that could interrupt conversations

  • Fixed a loop caused by active model fallback logic in chat

  • Fixed excessive re-rendering in the chat timer component

  • Fixed referral stats displaying NaN values

  • Fixed character selection overlay not dismissing properly on mobile

  • Fixed prompt history interfering with @mention keyboard navigation

  • Fixed library page getting stuck in a loading state

  • Improved authentication loading performance

  • Fixed theme color selector not indicating the selected color

  • Fixed promo codes not applying correctly during checkout

  • Fixed menu items appearing highlighted when not selected

  • Fixed selected theme not applying on the checkout page

  • Fixed typos in model descriptions for several models

  • Fixed incorrect error message when an invalid voice is specified

  • Fixed message order after editing by correcting the sorting logic

  • Fixed image modal footer visibility

  • Collapsed duplicated encrypted reasoning placeholder in TEE/E2EE responses

  • Fixed image modal orientation for large portrait images on mobile