Skip to main content
Open Playground in the console. Select a model in each panel, enter a system prompt, and send the same request through both. The left panel sends the request in Raw mode. The right panel sends it with SERV Reasoning. For each response, you can inspect the output, any reasoning returned by the provider, token usage, cost, and latency. Text works with every model. The attachment button appears only when both selected models support a common attachment type. Supported uploads are PNG, JPEG, or WebP images and PDF documents. Try a larger model in Raw mode against a smaller model with SERV. Test both with production prompts, edge cases, and incomplete inputs instead of judging them from one response. Open the SERV Playground →