OpenSketch is a free diagram and figure editor for scientific workflows to use in papers, presentations and posters. Here, you can combine scientific illustrations with editable shapes, labels and connectors, and export as formats such as SVG, PNG, PDF, PPTX and more. The app currently comes with a set of ~1000 biomedical SVG assets (detailed & simplified illustration modes) following a coherent design theme, all created with GPT-6 Astra.
What real task does your product handle with GPT-6 Astra?
Maker
GPT-6 Astra played a big part in development: I used it to create all ~1000 original pieces of artwork and help improve the editor’s performance and interface. Especially the former would not have been possible without it. It's really good at taking a reference PNG, create a new asset with the same styling, and then convert that PNG into SVG.
Report
Maker
📌
Hi everyone! I’m Paul, a PhD researcher working in cancer immunology.
I built OpenSketch for something every (biomedical) scientists frequently needs, which is creating clear scientific diagrams of experimental workflows or biological mechanisms for scientific presentations, posters, and articles.
Current solutions either 1) offer comprehensive asset sets but are locked behind paywalls, 2) are free but have incomplete asset sets, or 3) don't come with an editor environment. OpenSketch serves to resolve those issues by providing free, comprehensive asset sets, following consistent themeing, inside a performant editor. In addition, only the frontend is web-hosted; all personal project data is stored locally to ensure privacy and offline access.
GPT-6 Astra played a big part in development: I used it to create all ~1000 original pieces of artwork and help improve the editor’s performance and interface. Especially the former would not have been possible without it. It's really good at taking a reference PNG, create a new asset with the same styling, and then convert that PNG into SVG. Next, I also plan to generate asset sets for other fields, such as chemistry, physics, and more. If you have an area in mind, please let me know!
I’d especially love feedback from researchers, students and educators. Try making something you actually need for your work, and tell me where you get stuck. Which scientific illustrations are missing? What takes too many clicks? What would improve your user experience? Those are the things I’d like to improve next.
GPT-6 Astra played a big part in development: I used it to create all ~1000 original pieces of artwork and help improve the editor’s performance and interface. Especially the former would not have been possible without it. It's really good at taking a reference PNG, create a new asset with the same styling, and then convert that PNG into SVG.