What’s new.

We added five models: Flux 3 Video, MiniMax H3, Qwen Image 3, Qwen Image 3 Pro, and DeepSeek V4 Pro.

Resources:Flux 3 VideoH3Qwen Image 3Qwen Image 3 ProV4 Pro

CSuite now has a light theme. It follows your system setting by default, and you can pin it to light or dark in Account.

You can now trim video clips frame by frame, play them between half and double speed with the pitch kept, remove the audio or send it to the Audio workspace, and export a clip as a GIF.

The transport bar now has a loop toggle, preview speed, and single frame stepping with the arrow keys. You can save any frame as an image or use it as the start frame for a new generation.

An info popover now shows a clip's duration, resolution, frame rate, codecs, and how it was generated. From the AI menu you can upscale a clip to 1080p or 4K, and you can add a clip straight to a composition.

From the image toolbar you can now copy a picture straight to the clipboard and flip it horizontally or vertically. Resize also offers ready-made sizes for Open Graph, Instagram, Stories, and X that crop to fit rather than stretch.

Each image you generate now remembers how it was made. The info popover shows the model and prompt and can refill the composer with both, and your recent prompts sit as chips above it.

You can now adjust an image's brightness, contrast, and saturation with a live preview. The annotate tool also gained a brush, arrow, rectangle, text, and eraser, each step undoable.

A new AI menu on the image toolbar upscales two or four times and removes backgrounds. Generation now shows its progress in stages, and the composer shows roughly what an image will cost before you run it.

Audio now has a working format converter, edits that keep your file's format, an info popover with prompt reuse, prompt history, and a cost hint. With Fish Audio S2.1 Pro you can also clone a voice from a reference clip you pick in the app.

Resources:Fish Audio S2.1 Pro

Models is now split into your connected providers and a Defaults tab, where you set the model each workspace starts with. Those picks are shared with the model button everywhere else.

You can now play a sample of a text to speech voice from the picker before you generate.

We fixed local model setup on two systems. The HuggingFace runtime failed to install on Windows and Ollama setup could hang on Ubuntu, and both now complete.

We added a lot to the text editor: a slash menu for inserting blocks, search within a document, an outline for jumping between headings, live word and character counts, templates, export, and media you can embed inline. You can also attach files to a document as context for the AI, and the inline suggestions that write ahead of you got better.

Every generation workspace now opens on the lowest cost quality, resolution, and duration a model offers, so nothing expensive runs because you did not change a default.

We fixed missing inputs on Runware models. Video models now have the start frame, end frame, and settings their Replicate counterparts had, and image models now take reference and starting images.

Actions that used to fail quietly now show you a message explaining what went wrong.

We fixed Shorten, Expand, and Improve in the text editor, which only worked with local models and did nothing on a cloud provider.

You can now drop images, audio, or video onto a workspace sidebar to import them into your project folder, and drop media into a text document to embed it inline.

We rewrote onboarding. Setup is now six steps that end with your license: what CSuite is, connecting providers, what each workspace does, chat and workflows, and choosing a folder.

We added MiniMax Speech 2.8, Fish Audio S2.1 Pro, ACE-Step v1.5, Kimi K3, and Claude Sonnet 5. We also corrected a local model that pointed at a tag that did not exist, and fixed Nano Banana 2 Lite running the full model at roughly twice the cost.

Resources:Speech 2.8Fish Audio S2.1 ProACE-Step v1.5Claude Sonnet 5Nano Banana 2 Lite

We fixed a crash when you cancelled a text generation. It now stops cleanly.

CSuite now checks for new versions in the background, downloads them quietly, and installs them on your next quit. The Updates panel in Account shows the current status, lets you check manually, and has a switch to turn automatic updates off.

We fixed the window showing with no background at all on Linux setups without compositing.

This is the first public release of CSuite: text, images, audio, video, chat, and workflows in one place, running local models on your own machine or cloud models through your own provider keys. You can get it for macOS on both Apple silicon and Intel, for Windows, and for Linux as an AppImage.