Workshop Cloud + Desktop: One Platform, Two Surfaces
Workshop isn't confined to one surface. Cloud for quick builds in the browser, Desktop for deeper work on local files and existing projects. Both publish. Here's how they fit together.
TL;DR
Workshop comes in two forms: Cloud (browser-based, instant, zero setup) and Desktop (native app, local files, any framework). Both publish. They share the same AI agent and core capabilities, but they're optimized for different kinds of work. This post explains when to use which — and how they fit together.
Why Two Surfaces?
Most AI development tools make you pick a lane. Browser-based tools are fast to start but limited in what they can build. Desktop tools are powerful but require setup, and they usually can't deploy what you make.
We didn't want to choose. Different work calls for different environments:
- Prototyping a data app for your team? You want to start immediately, in a browser, and publish it with a link.
- Working on an existing codebase? You need local file access, your own dev environment, and the freedom to use any framework.
- Building something private with sensitive data? You want everything to stay on your machine.
Workshop is one platform that spans both. The AI agent is the same. The way you work — Plan mode, Build mode, connectors, checkpoints — transfers directly. The difference is where the work happens.
Workshop Cloud
Workshop Cloud runs in your browser. No installation. No setup. Sign up and start building.
What It's Built For
- Quick builds — go from idea to working app in a conversation
- Publishing — one-click deploy to a live URL with auth and access control
- Collaboration — share projects with teammates, iterate together
- Data apps and dashboards — connect to your databases, build interactive tools, share with stakeholders
What You Get
| Feature | Details |
|---|---|
| Runtime | Managed sandbox (up to 4 CPU, 16 GB RAM) |
| Tech stack | HTML + Python (FastAPI), Streamlit |
| Publishing | One-click deploy with shareable URL |
| AI connectors | Managed (no keys) + BYOK |
| Data connectors | PostgreSQL, BigQuery, Snowflake, MongoDB, and more |
| GitHub | Link and sync projects |
| Live preview | Built-in preview pane as you build |
Cloud is the fastest path from intent to shipped outcome. If what you're building can live in the browser and you want to share it with a URL, start here.
Best For
- Data apps, dashboards, and internal tools
- AI-powered web apps with managed model connectors
- Websites you want deployed fast
- Teams that need a shared, accessible environment
Workshop Desktop
Workshop Desktop is a native application for macOS, Windows, and Linux. It runs on your machine, works with your files, and has no session limits.
What It's Built For
- Existing projects — open a repo, understand the code, make changes, run tests
- Deep work — no session timeouts, no compute limits, your full local environment
- Privacy — local models, local data, nothing leaves your machine
- Any stack — not limited to Python/HTML — use any language, any framework
What You Get
| Feature | Details |
|---|---|
| Runtime | Your local machine (unlimited) |
| Tech stack | Any language, any framework |
| Local models | Connect to llama.cpp, Ollama, or any compatible server |
| MCP servers | Built-in directory + custom integrations |
| File system | Full local file access |
| Session limits | None |
| Deployment | Deploy anywhere — Vercel, AWS, Docker, your own server |
Desktop is for work that needs more room. Bigger projects, sensitive data, custom toolchains, or the kind of extended sessions where you don't want to think about timeouts.
Best For
- Working on existing codebases and repos
- Projects that need specific frameworks (React, Swift, Rust, etc.)
- Private work with local models — zero cloud calls
- Long development sessions without interruption
- Deploying to your own infrastructure
How They Fit Together
Cloud and Desktop aren't competing products. They're two surfaces of the same platform, and the typical workflow uses both.
Start in Cloud, Continue on Desktop
The most common pattern:
- Prototype in Cloud — quickly build a working app, dashboard, or website in the browser
- Push to GitHub — Cloud has built-in GitHub integration
- Open on Desktop — clone the repo and continue with full local control
- Iterate locally — add features that need a different framework, local data, or local models
- Deploy on your terms — push to Vercel, AWS, or wherever you host
This means your work isn't locked into one environment. The fast-start convenience of Cloud and the deep flexibility of Desktop complement each other.
Use Both at the Same Time
Some teams use Cloud for data apps and dashboards (where publishing and sharing are key) and Desktop for heavier engineering work (where local files and custom stacks matter). Same AI agent, same account, different surfaces.
What They Share
A system-aware agent that works end-to-end Workshop's agent understands the full workflow — what to build, which connectors to use, and how to get things ready to ship. It works across Cloud and Desktop, powered by frontier models or local models.
A real runtime environment The agent works inside a real execution environment. In Cloud, that environment runs in the cloud. In Desktop, it runs on your machine, with access to local files, runtimes, and your existing tools. The stack isn't preset — Workshop builds across whatever tech fits the project.
Managed connections to AI and your systems Workshop handles model selection, credentials, billing, and rate limits through managed connectors for Anthropic, OpenAI, and Gemini — or bring your own keys. Beyond AI, Workshop connects to databases, warehouses, GitHub, and the tools your workflows depend on.
Built-in publishing and sharing When you're ready to ship, Workshop handles hosting — no separate deployment project needed. Publishing includes access control: share publicly or with specific people, as part of the same flow.
Which Should I Choose?
Not sure? Start with Cloud. It's the fastest path from idea to working app. If you hit a limitation — need a different framework, want local models, or need to deploy to specific infrastructure — move to Desktop. The transition is smooth because the AI agent works the same way.
| I want to... | Use |
|---|---|
| Build a data app or dashboard fast | Cloud |
| Publish without any local setup | Cloud |
| Work on an existing codebase | Desktop |
| Use local AI models (private, offline) | Desktop |
| Build with React, Swift, Rust, etc. | Desktop |
| Deploy to my own infrastructure | Desktop |
| Prototype quickly, then go deeper | Cloud → Desktop |
Try Workshop Cloud — free, no credit card required.
Download Workshop Desktop — macOS, Windows, and Linux.
For a detailed comparison, see Cloud vs Desktop.
FAQ
Do I need both Cloud and Desktop? No. Either one works on its own. Cloud is great for quick builds and publishing. Desktop is great for deeper work. Many users start with one and add the other when they need it.
Can I move a project from Cloud to Desktop? Yes. Push your Cloud project to GitHub, then clone it on Desktop. Your code, structure, and context transfer cleanly.
Does my account work across both? Yes. One Workshop account works on both Cloud and Desktop. Your connectors, settings, and subscription carry over.
Are local models available on Cloud? No. Local models are Desktop-only by design — they run on your hardware for privacy and offline capability. Cloud uses managed or BYOK cloud models.
Is Desktop free? Workshop Desktop is free to download and use. Local AI models are unlimited and free. Cloud model usage (Anthropic, OpenAI, Gemini) is billed through your Workshop subscription or credits.