... most-requested feature is live! Introducing the Prompt Playground . Write one prompt, pick your models (GPT, Claude, Gemini, Flux, etc...), and compare outputs side by side. No more tab-switching. Works with your own API keys. Especially powerful for image generation ( see the video example running a simple prompt about the World Cup through Flux, Imagen, and Gemini Flash). Obsessed with this one Open to your thoughts - we're moving beyond just improving prompts, to become a proper AI toolkit ...

... been experimenting with ways to streamline API testing without switching between Postman and VS Code. To explore this, I built a small extension that lets you send requests, debug responses, and visualize data directly in the editor. Over the past couple of months, I ve seen some interesting adoption numbers (300+ acquisitions with ~100 page views), which made me wonder: How do you all handle API testing in your workflow? Do you prefer dedicated tools like Postman/Insomnia ... ... integrate testing directly into your editor/IDE? What features would make an in editor API tool genuinely useful for you? For anyone curious, here s the Marketplace listing: VS

... Your AI editor just got superpowers for distributed systems design. I've shipped a remote MCP server that connects Claude, Cursor, VS Code, OpenCode, and Windsurf directly to your Cascode account over OAuth 2.1. No API keys. No copy-paste. Just your AI editor talking natively to your diagrams. **What's new in 0.8.0:** **Proper auth, done right** OAuth 2.1 with PKCE and dynamic client registration. Per-user token scoping and audience-bound access tokens mean your diagrams stay yours ... ... tweak it, and push it back. **Terraform CloudFormation round-trips** this is the big one. Convert between formats while preserving every node and edge. Infrastructure-as-code finally has a format-agnostic home. **One-click install for Cursor and VS

... siloed SaaS with one unified credit wallet and seamless cross-app workflows. It lets users automate video editing, content repurposing, CRM updates, client billing, lead scraping, and more via three simple build modes: AI-prompt description, step-by-step Playground preview, or full visual canvas saving 10-20 hours/week while cutting tool costs 60-80%. Core Features Unified Apps : 25+ built-ins (YouTube tools, blog generators, resume builders, CRM, calendars, media suites) sharing auth, credits, and state no API keys ... ... external logins. Workflow Building : Prompt AI to assemble recipes (e.g., "YouTube clips blog social"), run in Playground

... Introducing Playground , a full real-time GLSL development environment within DASCA for writing shaders and animating them with advanced curve support, alongside the new Depth Ball effect and general performance improvements ...

... team of 16 people in a use-it-or-loose-it way. For a tool we use maybe 2 3 times per sprint. That s just ridiculous! IMO, it should be something more like JSFiddle. A simple prompt playground that does not require you to signup, does not require API keys, and let's experiment instantly, i.e. you just enter a browser URL and start working. Like JSFiddle has. And mainly, something that costs me nothing ... ... measure awareness (visits, comments, reactions) and initial activation potential (people actually running prompts) before investing further. Setup at launch At launch I had: No landing page. Just a sidebar saying LangFast is your go-to prompt playground

... Same input. Same config. Totally different output. Because LLMs aren t deterministic and when your product depends on stable responses, it feels like building on sand. I got tired of playing prompt roulette. Tired of: Copy-pasting prompts into playgrounds Switching models manually Changing temperature settings Scoring responses by eye Losing track of what worked and what didn t So I built PromptPerf A fast, focused prompt testing playground