Hi everyone, just a heads up that Avcs is a local-first AI Visual Content Studio built around Codex.

Avcs started from a practical need: I already had a ChatGPT Pro subscription and did not want to pay for another dedicated image generation subscription just to keep making visuals. So Avcs uses codex app-server to turn Codex into a local image workflow.

One of the coolest parts is that image generation does not have to start from prompts alone. Avcs can load accurate reference assets from external data providers, keep imported and generated images in local project folders, and let you reuse any result as a reference in the next Codex turn.

It also works close to your real project files. You can ask Avcs to understand a local project, generate screenshots, posters, banners, annotations, or other visual content, and collect everything on a freeform Output board.

This project uses the local-first app architecture from livebook-dev/elixirkit .

Enjoy!

Project:

Download for macOS Apple Silicon:

Please download the file named Avcs-macos-aarch64.app.zip.

Note: The app is not notarized.