SurgePix: AI visuals. Edit every layer. Agent-callable.

The agent skills live in a public GitHub repository and cover background removal, slide deck generation, image translation and region-level inpainting. In one documented example, an instruction to make an awning blue changes only that region and leaves the rest of the image untouched. A separate text extraction feature splits text in an image into editable layers, so copy can change without re-rendering the artwork. The skills work with Claude Code, Codex CLI, Cursor, Gemini CLI, and any agent that supports the SKILL.md standard. Every skill also runs in a browser for people who prefer a visual interface. The product's public launch was set for October 26 on Product Hunt. Product Hunt's AI generative media category describes the company as emphasizing editable, brand-safe image generation with layer control and API access, and places it alongside tools such as Midjourney and RunwayML, which are better known for prompt-to-image and video generation. The company also says its Designer Agent searches the internet, reads pages, then generates visuals from what it finds, which goes beyond a model's training data. For a team building an AI company, the notable choice is that the visual step is exposed as something an agent can call, revise and chain together. That makes image generation a component inside a larger pipeline, such as a marketing workflow that produces assets and then localizes them. Before depending on it, founders should test output consistency across revisions, check the terms governing generated assets, and verify the claims in the launch materials, since most of the capability details come from the company itself.