Voxta v1.8.0: Living Avatars, Scenario Superpowers & a Fresh New Look

Voxta v1.8.0 is here - and it's a big one. This is really two releases in one: everything from our 1.7 line rolled together with a huge 1.8 drop on top. It's built around three big ideas: characters you can actually see and animate with full VRM support, a massive scenario authoring and scripting upgrade that turns scenes into living, reactive experiences, and a fresh coat of paint across the whole app - including a brand-new light theme and pick-your-own accent colors.

There's a lot here, so grab a coffee.

🎭 Living Avatars - VRM & Animation

Your characters can now stand on stage and move.

• Full VRM avatar support with 4 state animations, and multiple VRM characters on stage at once

• VroidHub integration - browse, import, and even create a character straight from VroidHub

• VRM animation preview studio to line up and test motions

• Play HY-Motion SMPL clips directly on the in-app VRM avatar

• Unity Asset Store animation support for VRM

• New /animate and /perform chat commands

• Transparent VRM pop-out in Voxta Desktop - float your character over anything

• A female mannequin for animation testing, alongside continued VRM display, hands, eyes and behavior polish

📝 Scenario Authoring, Leveled Up

Building scenes is a different experience now.

• An AI assistant right inside the scenario editor - now overhauled with persistent threads, image history, rollback, message actions, and per-change Apply or Reject

• Events, actions, buttons and contexts have been merged into one system, with a new graph view to see how it all connects

• New event triggers with counters and sequential events

• Scenario initial flags, plus "group by enum flags" to keep large scenarios organized

• Prompt Formatting Editor and a dedicated Assistant configuration

🎬 Scripting & Cinematics

Scenario scripts can now drive the whole experience.

• A player-facing HUD driven entirely by your scripts, with script-controlled visibility

• Real-time timers with chat.setTimeout and chat.setInterval

• Completion callbacks for script messages, and silent inference with chat.setCharacterCanSpeak

• Cinematic effects - flashes, shakes, glows, tints and more

• Video playlists for backgrounds and avatars with shuffle, sequence, and one-shot playback (with seamless looping)

• The Choice stage popup for modal player decisions

• Command mode - just say "enter command mode" for a spoken command palette and offline indicator

🎨 A Fresh Look & Better UX

• Brand new light theme with a Dark/Light toggle, plus user-selectable accent colours across the app

• A system monitor showing GPU, CPU and memory usage per module

• Login with Voxta Cloud directly - no more copy-pasting API keys

• Redesigned model selection field and visual slot cards for picking scenarios and characters

• A Modules testing page to verify every installed module at once

• Assets explorer and direct upload in the Characters/Scenarios assets tab

• Persistent image, speech, and animation playground histories with restore and delete

• Curated voices support with a much better voices browser

• Per-chat Narrator toggle that skips narrator TTS when muted

• Background removal in the Image Generation Playground

🧩 New Modules & Backends

• OmniVoice TTS and SenseVoice STT (which can read some noises and emotions)

• Krea 2 Image Generation and a Qwen Image Edit module for image-to-image

• A new Claude Code TextGen module with model, reasoning, image and large-context support

• Diffusers-based background removal

• ExLlamaV3 updated to 1.1 with vision model support, KV cache quantization, speculative decoding, banned strings, loop detection, and a fix for the 16k context cap

📦 VoxPak & Packaging

• A brand new VoxPak package format with deduplication, better compression, and runtime asset streaming without extraction for published content

• Support for very large content packages

• Google Drive integration for content backup (per-item)

• Packaging progress shown while exporting resources

⚙️ Platform & Installer

• Upgraded to Python 3.13, Torch 2.9.1, and Transformers 5 (this needs a full reinstall of Python and packages)

• ARM64 Linux build

• Gemma 4 prompt template, plus built-in support for thinking templates and prefixes

• Greatly improved built-in and semantic-kernel memory, and a chat journal that rewinds flags and variables too

• Installer can now pick which existing Voxta install to update or repair, with validation to protect unrelated folders

• Voxta Mobile can now upgrade in place thanks to a stable signing key

...plus a long list of stability and polish fixes across the app.

⚠️ Heads up: because this release upgrades Python and Torch, if you use local models your next Install will download a few GB of content. Signed releases are also live now, so no more skipping Windows security on install.

Thank you all for the continued support and feedback that makes releases like this possible. Whether you're posing VRM characters, scripting cinematic scenes, or just enjoying the new light theme - we hope you love v1.8.0.

After subscribing to the required Patreon tier, you can get the latest version at: https://portal.voxta.ai/