Vresk launches AI workspace for side-by-side model comparisons

Vresk opened its web-based AI workspace to the public this week at vresk.ai and launched on Product Hunt today. The product lets users compare up to three open AI models at once, while keeping notes, files and memory in one workspace across models.

Why it matters: - Vresk is targeting a common pain point for people using multiple AI tools: repeating the same prompt, re-uploading files and rebuilding context in different apps. - The workspace is built around open-weight models from the United States, Europe and China, which are arriving at a steady pace and often perform differently on the same task. - The product also gives users a way to see which model produced each answer, which can make comparisons faster and more transparent.

What happened: - Vresk opened to the public this week at vresk.ai and launched on Product Hunt today. - The web-based workspace lets a user send one prompt to up to three open AI models at once. - The replies appear side by side, with each answer labeled by the model that wrote it and a badge showing where that model was built. - Users can continue working with any model on the shelf after the comparison.

The details: - Notes, uploaded files and memory stay in the user's workspace instead of being tied to one model. - Vresk includes a label on every response that names the model behind the answer. - A single setting limits the available models to those built in the United States and Europe. - Users can view what Vresk remembers about them, turn memory off or delete it. - The same chat can generate interactive tools, including a mortgage calculator. - The workspace can also create designed slide decks that export to PowerPoint. - Vresk is available in any modern web browser. - A free tier does not require a credit card. - Guest access is available without creating an account. - Vresk Plus costs US$20 per month at a flat rate and adds the largest models on the shelf.

Between the lines: - Vresk is betting that model choice will matter more as the number of capable open models grows. - Keeping memory and files inside a user workspace rather than inside one model reduces switching costs and could make side-by-side testing more practical. - The origin badge and model label signal a focus on accountability as well as comparison. - Founder Josh Hikida framed the product as a way to test new models on a user's own work before committing to one.

What's next: - Vresk is positioning the product for broader use through browser-based access and a free entry tier. - The launch on Product Hunt could help drive early adoption and feedback from people already comparing AI models. - The company is likely to compete on workflow convenience as more open models reach the market.

The bottom line: - Vresk wants to make open-model comparison feel like one workspace, not a stack of separate apps.