VidTSX Studio 1.1.0

The first real release. Everything below runs on your own machine with your
own provider keys or a Claude subscription. No account, no server in the
middle, no telemetry (crash reports only if you say yes, see below).

What's new since 1.0.0

Studio, an AI-assisted video editor. Import your clips (4K HEVC works;
edits run on proxies), transcribe them, and let Auto Cut remove silences,
retakes and fillers into a normal timeline you can split, trim and ripple by
hand. An editing assistant in every project takes "edit this video end to
end", transcribes, cuts, proposes TSX shots and b-roll, turns on captions and
queues the export, with every change arriving as a card you accept or reject.
Brands, editing presets that learn from how you cut, word-synced captions,
per-clip transitions and filters (with importable packs), a Fast export
engine that copies untouched footage straight from the source, and a
.vidtsx package for moving a project between machines. Also new: SVG
images in the media pool, a Notes tab per project the assistant can read,
six optional extra brand colours, brand fonts that load in the preview and
in the export, and Import TSX that bundles a multi-file composition into one
shot and attaches the images it uses (swap them in the shot's Media
section).

Videos. Generate clips from text, an image or a reference video with
Seedance on fal or BytePlus ModelArk, or fully offline with Wan through the
bundled runtime.

Agents. Five built-in agents (Assistant, Motion Post, TSX Composer, Web
Designer, Flow Builder) that take a brief to a finished result inside the
app, plus installable .vidtsxagent packages.

Flows. Node-based recipes you run yourself or hand to an agent. Five
built-ins ship; the editor builds more.

Templates. A third mode in the TSX creator: ten ready-made compositions
with a form instead of a prompt, including three transparent overlays.
Templates can carry sound (a field to pick, play or silence each one), and
more templates install from .vidtsxtemplate files (Import, or double-click
the file).

Home, project posters, an in-app model picker everywhere an LLM runs, a
3D Studio (image to 3D), background removal, and a content-safety gate on
every generator.

Also: local image models are a folder-as-truth library you can drop
files into; Cloudflare Workers AI, Google and OpenAI subscription image
providers; ElevenLabs transcription; encrypted key storage; the app updates
itself.

Crash reports, opt-in. On first launch the app asks once whether it may
send crash reports (stack trace, app version and OS; never your prompts,
files or keys). Off unless you say yes; change it anytime in Settings,
Privacy.

Providers in this release

• LLM: Claude (subscription or API key, including Fable 5.1, Opus 5.5 and
  Sonnet 5.5) and OpenRouter.
• Images: Fal, BytePlus ModelArk, OpenRouter, Cloudflare Workers AI, local
  models, and the Google and OpenAI CLI subscriptions.
• Video: Fal, BytePlus ModelArk, local.
• Transcription: local whisper, AssemblyAI, ElevenLabs, OpenRouter.
• Sound effects and music: ElevenLabs (through the Studio assistant and
  agents).

More LLM providers return in a later release once the current set has had
feedback.

Requirements

Windows 10 or 11, x64. A Vulkan-capable GPU is recommended for local image
and video models.

Known issues

• The installer is not code-signed yet. Windows SmartScreen shows a warning
  on first launch; choose "More info", then "Run anyway". Updates install
  silently afterwards.
• If you reload the app while a video generation is running, its progress
  card disappears. The job keeps running and the clip still lands in the
  library.
• Cost figures in the usage dashboard are estimates at list rate for most
  models.

Updating from 1.0.0

Nothing to do. The app checks for updates on launch and installs this one on
the next restart.